You are on page 1of 21

PDHengineer.

com
Course № E-3022

Power System Neutral Grounding


Fundamentals

To receive credit for this course

This document is the course text. You may review this material at
your leisure either before or after you purchase the course. To
purchase this course, click on the course overview page:

http://www.pdhengineer.com/pages/E-3022.htm

or type the link into your browser. Next, click on the Take Quiz button
at the bottom of the course overview page. If you already have an
account, log in to purchase the course. If you do not have a
PDHengineer.com account, click the New User Sign Up link to create
your account.

After logging in and purchasing the course, you can take the online
quiz immediately or you can wait until another day if you have not yet
reviewed the course text. When you complete the online quiz, your
score will automatically be calculated. If you receive a passing score,
you may instantly download your certificate of completion. If you do
not pass on your first try, you can retake the quiz as many times as
needed by simply logging into your PDHengineer.com account and
clicking on the link Courses Purchased But Not Completed.

If you have any questions, please call us toll-free at 877 500-7145.

PDHengineer.com
5870 Highway 6 North, Suite 310
Houston, TX 77084
Toll Free: 877 500-7145
administrator@PDHengineer.com
Power System Neutral Grounding Fundamentals

by

Louie J. Powell, PE
Saratoga Springs, NY

One of the decisions that must be made in designing a power system is how the neutral of the
system should be connected to ground. The designer has several options from which to choose:
ƒ No intentional connection between the neutral and ground
ƒ Solid (no intentional impedance) connection between neutral and ground
ƒ Insertion of resistance between neutral and ground
ƒ Insertion of inductive reactance between neutral and ground
And within the resistance and inductive choices, the designer has a further decision regarding the
relative magnitude of the impedance to be inserted into the circuit. Ultimately, each of these
choices affects the way that the power system performs in response to various contingencies, so
the choice between these options is not a trivial matter.

It is possible to devote an entire career to studying the implications of these choices. Recently,
three noted authors published a nearly 600 page book1 that addresses the concerns involved in
grounding decisions in great detail. This seminal book should be included in every serious library
treating power system engineering fundamentals. IEEE has published a number of standard
references on the subject that should be available to every power system engineer.2 3

But it is also possible to visualize the impact of the basic choices in an intuitive fashion that does
not rely on heavy use of mathematics. This course will present the basic choices as well as the
resulting system performance characteristics.

Introduction

Figure 1 illustrates a simplified power system consisting of a three-phase voltage source


connected to a set of conductors. The three-phase source consists of three Thevenin equivalent
fundamental frequency sinusoidal voltage sources that are each equal in magnitude to the phase-
to-ground system voltage, and that are displaced 120 electrical degrees. The electrical system is
represented by a set of three inductive reactances, designated as jXL, that are equal in magnitude.
There may be some unbalance on practical power systems, but the impact of unbalance between
phases is beyond the scope of the present treatment.

Practical power systems also include resistance, but in most instances the inductive reactance is
about an order of magnitude larger than the resistance. So for the sake of this development,
system resistance can be ignored.

Attention is drawn to two important aspects of Figure 1. First, two terminal points have been
identified in the figure. N is a terminal point at the neutral of the power system – the neutral of
the three-phase voltage source. G is a terminal point that is connected to ground. In this context,
“ground” is a reference plane that is connected to earth (in Europe, the term “earthing” is used to
convey the same meaning as the term “grounding” in North America) and that is the reference
point for all voltages throughout the system. Those two terminal points will be retained as the
figure is later transformed into an equivalent circuit for analysis so that it is possible to explore
the impact of the various choices of how N and G may be interconnected.

© 2009 Louie J. Powell page 1 of 20 pages


Power System Neutral Grounding Fundamentals

Figure 1 also shows that there is a capacitance, shown here as capacitive reactance, -jXC0,
between each phase conductor and ground. This is a distributed capacitance that exists by virtue
of the fact that the power system conductor is physically in parallel with earth. The negative sign
indicates that this distributed parameter is capacitive. The suffix 0 has a meaning drawn from the
study of symmetrical components. However, it is not necessary for the reader to understand
symmetrical components, and it is sufficient to simply accept that the suffix is a convenient
handle to assign to the distributed parameter that may reappear later in some other aspect of
power system engineering analysis.

Fig 1 Typical three-phase power system with system inductive reactance (JXL) and distributed capacitance (-JXC0), and
with terminal points that can subsequently be used to explore options for connecting the system neutral (N) to ground
(G).

System inductive reactance in ohms can be calculated from the inductance of the system using
equation 1:
jXL = j 2π f L (1)

where
f is system frequency
L is the inductance in the system in henries

Likewise, the distributed capacitive reactance (also in ohms) can be calculated using equation 2.

1
− jXC 0 = (2)
j 2π f C
where
f is system frequency
C is the distributed capacitance to ground in Farads

Readers will recognize that the capacitive reactance looks a bit like a load. If fact, this capacitive
reactance is a parasitic charging capacitance through which current is always flowing. The only

© 2009 Louie J. Powell page 2 of 20 pages


Power System Neutral Grounding Fundamentals

reason it is normally not considered is that –jXC0 is very much larger than jXL – several orders of
magnitude larger – so it is normally negligible.

Because –jXC0 is much larger than jXL, it really doesn’t matter where the distributed capacitors are
actually connected in the equivalent circuit. In fact, we can easily move the connection point
from where it is shown in figure 1 across the series inductive reactance to a set of points between
the three voltage sources and their associated reactances. Having made that change, and
recognizing that the voltage source is a Thevenin equivalent voltage and therefore impedanceless,
we can further move those shunt capacitance elements down to the neutral of the voltage source.
However, when that is done, the distributed capacitive reactances of the three phases are in
parallel, and can be replaced with an equivalent distributed capacitance value of –jXC0/3 as shown
in figure 2.

Fig 2 – Equivalent circuit with the distributed capacitance lumped at the neutral

Note that in figure 2, terminals N and G have been retained.

Finally, we can take one further step in the evolution of this equivalent circuit by observing that
figure 2 is totally symmetrical – it represents three phases that are equal in magnitude and
displaced from each other by 120 electrical degrees. Therefore, for the sake of analysis, we can
ignore two of those three phases, perform our analysis on the third phase only, and then observe
that whatever we find happening on that phase will also happen on the other two phases 120o
(approximately 5.555 msec) and 240o (about 11.11 msec) later, respectively. That then leads us
to the simple, single phase equivalent circuit shown in figure 3. And once again, we observe that
the neutral, N, and ground, G, terminals have been retained.

So now the question is: considering the simplified equivalent circuit of figure 3, what is the
consequence of the following options:
1. Open circuit between N and G
2. Zero-impedance connection between N and G
3. Insertion of a low magnitude of resistance between N and G
4. Insertion of a high magnitude of resistance between N and G
5. Insertion of a low magnitude of inductive reactance between N and G
6. Insertion of a high magnitude of inductive reactance between N and G

© 2009 Louie J. Powell page 3 of 20 pages


Power System Neutral Grounding Fundamentals

Fig 3 Single phase equivalent circuit for a three-phase power system

In order to investigate those six options, however, we do need to make one final modification to
the equivalent circuit. The basic problem that the options present relates to what happens when
there is a single-phase-to-ground fault on the power system, and to represent such a fault, we need
a switch. Hence, the final equivalent circuit is shown in figure 3. Closing this switch has the
effect of applying a single-line-to-ground fault on the system. Leaving the switch open is
equivalent to having the system unfaulted.

Fig 4 Single-phase equivalent circuit for investigating system neutral grounding options

Case 1: Open circuit between N and G

This case represents the system application in which there is no intentional connection between N
and G; that is, the system is nominally “ungrounded”. In reality, of course, the term
“ungrounded” is inexact because the neutral is really connected to ground through the reactance
of the distributed charging capacitance in the system. That is, an “ungrounded” system is actually
capacitively grounded through –jXC0/3.

© 2009 Louie J. Powell page 4 of 20 pages


Power System Neutral Grounding Fundamentals

Because the impedance that actually limits ground fault current is rather high, the magnitude of
current that will flow to a ground fault is often so low that automatic tripping is not required.
That is the most-cited advantage of the ungrounded system – the fact that ground fault tripping is
not required means that system availability can be attractively high. This is especially appealing
in mission critical applications such refinery and power house auxiliary systems

Earlier it was noted that –jXC0 is several orders of magnitude larger than jXL. Therefore, it also
must be true that –jXC0/3 is much larger than jXL. For this reason, jXL can be ignored, and this
case then becomes a matter of capacitor switching.

When the switch is open, the voltage across the switch is equal to the single-line-to-ground
voltage on the system. But that voltage is also the same as the voltage source, V1. Therefore,
when the switch is open, the voltage across the equivalent distributed capacitance, –jXC0, must be
zero (Kirchoff’s voltage law requires that the voltage across the capacitor plus the voltage V1
must equal the open circuit voltage.)

Closing the switch (applying the ground fault) forces the voltage across the switch to be zero.
Therefore, the voltage across the capacitor must be –V1 (again, Kirchoff’s voltage law). But
principle that most of us remember from fundamental physics is that it is not possible to change
the voltage across a capacitor instantaneously. It’s not possible for the voltage across the
equivalent distributed capacitance to change from zero to –V1 instantaneously!

So that suggests another principle that we may remember from many years ago – the switching of
an energy storage element (such as a capacitance) always involves a differential equation, and the
solution to that differential equation always includes two components:
ƒ A steady-state component
ƒ A transient component
And the nature of these components depends on when the switching occurs.

Recall that V1 was described as a fundamental frequency sinusoidal voltage. If the instantaneous
value of V1 at the instant the switch is closed is zero, then the instantaneous voltage across the
open switch in the instant just prior to it being closed is also zero. Because the voltage across the
switch must remain zero after it is closed, then the voltage across the distributed capacitance will
gradually increase to a value of –V1 at the same time that the driving voltage in the network
increases to V1.

The steady state component of current will be the voltage divided by the magnitude of the
distributed capacitive reactance, while the transient component will be zero. Since the distributed
capacitive reactance is very large, then the steady state current (ground fault current) will be quite
small – essentially equal to the nominal capacitive charging current that would flow when the
system is otherwise healthy.

But if V1 is at its crest value at the instant the switch is closed, then a transient component must
be generated by the switching event. That is –
ƒ Prior to closing the switch, the voltage across the switch is V1 and is at its crest value
o The voltage source is also V1 and is also at its crest
o Therefore the voltage across the equivalent capacitance must be zero
ƒ But in the instant after closing the switch, the voltage across the switch is zero
o The voltage source continues to be V1 and continues to be at crest.
o The voltage across the equivalent capacitor must continue to be zero since it
cannot change instantaneously.

© 2009 Louie J. Powell page 5 of 20 pages


Power System Neutral Grounding Fundamentals

ƒ Therefore, a transient voltage of magnitude –V1 (crest) must appear in the circuit in the
instant after the switch closes.

The implications of these observations are that the application of a ground fault in a power system
with the neutral ungrounded (which really means grounded through distributed capacitance) will
result in generation of transient overvoltages. The magnitude of the transient overvoltage
depends on both system parameters (obviously on capacitance, but also on inductance and
resistance) and on when during the sinusoid of voltage on the AC system the fault occurs. One
might imagine that faults are more likely to occur when the AC system voltage is nearing its crest
value (because that implies the maximum stress on insulation), and unfortunately, that is exactly
the condition that results in the most severe transient overvoltage. The phenomenon of excessive
transient overvoltages due to ground faults on nominally ‘ungrounded’ systems has most often
been observed in higher voltage applications.

After the fault has been on the system for a period of time the transient overvoltage will decay
away leaving a voltage across the capacitor that is equal to –V1 (because the voltage across a
closed switch must be zero). If the fault then clears, the current required to charge the distributed
capacitance will be discontinued, leaving the voltage across the capacitance at the value of –V1
at the instant the fault was removed. Obviously, if this happens when V1 is at or near the crest of
the sinusoid, then the capacitance will be left charged at this steady state magnitude. Then, if the
fault is subsequently reapplied, depending again on when on the sinusoid of the source V1 the
fault reoccurs, the transient voltage that will have to appear in the circuit can be as high as 3V1.

This repeated cycle of interruptions and reignitions would normally caused by an external means
– say, vibration due to the rotation of a motor. The consequence of so-called ‘repetitive
restriking’ is to cause the sustained line-to-ground voltage to “ratchet up” to a value that is
significantly greater than V1. Repetitive restriking has most often been observed in low-voltage
systems.

There is one very special application of the ungrounded neutral that should be mentioned –
marine systems. The concern in marine systems is for cathodic corrosion of the hull of vessels,
and one way to manage that risk is to completely separate earth ground from the electrical
system. This creates special problems in system design and protection that are beyond the scope
of this course.

Finally, the supposed advantage of the ungrounded system – that automatic tripping for single-
line-to-ground faults is not required – is itself a potential problem. While the magnitude of
current for a ground fault may not necessitate tripping, the fact is that during a ground fault an
abnormally high voltage will be present on the two ‘healthy’ phases, and that phenomenon may in
fact accelerate the occurrence of a second fault. And with a pre-existing ground fault on the
system, a second ground fault would create a line-to-line fault.

Conclusions: “Un-grounded” systems


ƒ Very low ground fault current
o Essentially equal to the nominal system charging current
o Generally not detectable with ordinary current-based ground fault protection
technology
o May not necessitate automatic tripping for ground faults
ƒ Potential for hazardous transient overvoltages
ƒ Potential for hazardous sustained overvoltages due to repetitive restriking in low voltage
systems

© 2009 Louie J. Powell page 6 of 20 pages


Power System Neutral Grounding Fundamentals

ƒ Ungrounded marine systems require special consideration


ƒ Potential for accelerated second fault if the first fault is not automatically cleared.

Case 2: Zero-impedance connection between N and G

Applying a zero-impedance shunt between N and G has the effect of shorting out the distributed
capacitance of the system. Effectively, this is what would be accomplished by ‘solidly
grounding’ the system by connecting the neutral directly to ground with no intervening
impedance. Obviously, the transient effects associated with system capacitance are no longer a
concern (although there are other sources of transient overvoltages that would have to be
addressed).

However, there is a downside – if –jXC0 is removed from the circuit, the only impedance left to
control the magnitude of ground fault current is jXL. That means that that the magnitude of
ground fault current in the system would dramatically increase, and it would be the high
magnitude of fault current itself that would be troublesome. Specifically:
ƒ The single-phase-to-ground-fault current magnitude would be about the same as the
three-phase fault magnitude.
ƒ It is possible for the fault current for the current available to single-line-to-ground-faults
to greatly exceed the available three-phase magnitude. This phenomenon would most
likely be seen in association with solidly grounded generators or in close electrical
proximity to banks of wye-connected transformers.
ƒ Higher fault current implies greater stress on the circuit breakers that would have to
interrupt those fault currents.
ƒ Higher fault currents implies greater burning damage at the point of a fault. This is an
especially important consideration with rotating machines. A stator ground fault in a
motor or generator on a system where the neutral is solidly grounded will produce so
much burning damage that the machine will likely be unrepairable.
ƒ Higher fault current implies greater arc-flash hazard for employees in the workplace.
This subject has gotten a lot of attention in recent years. The situation where the arc flash
energy available in the system is so high that appropriate personnel protection clothing is
simply not available would likely be viewed as completely unacceptable by many
employers.
ƒ Higher ground fault current implies higher earth potential gradients. While this
phenomenon is not encountered often, it can be very disruptive for ground faults to
expose workers to distracting and potentially hazardous electric shocks even when they
are not in the immediate proximity of faulted electrical apparatus.
ƒ While these adverse consequences are a consideration, there are applications where the
higher fault ground fault current associated with a solidly grounded neutral has
advantages. Specifically, higher fault currents are usually desirable on high-voltage
transmission lines to be able to provide reliable protection over the entire length of a line.

The term “solidly grounded” is generically descriptive and applies to the situation where the
connection between N and G does not involve any intentional impedance. However, that term
does not provide any guidance on the resulting system performance. The term “effective
grounding’ has been defined to apply specific technical criteria to the conditions that will exist
when the neutral is solidly grounded (equations 3 and 4)4. Systems that are effectively grounded
will have known performance characteristics with respect to transient overvoltages, and whether
or not a system is effectively grounded is a criterion in the application of surge arresters. Most
solidly grounded systems are effectively grounded. But it is possible for a solidly grounded

© 2009 Louie J. Powell page 7 of 20 pages


Power System Neutral Grounding Fundamentals

system involving very long transmission or distribution circuits to not be effectively grounded,
especially at the remote ends of those circuits.

Criteria for effective grounding:


X0 ≤3 (3)
X1
R0 ≤1 (4)
X1

In these expressions,
X0 is the zero sequence inductive reactance
X1 is the positive sequence inductive reactance
R0 is the zero sequence resistance

Likewise, there are a few special instances in which effectively grounded systems are not solidly
grounded. Those will be discussed later in case 5.

There is one set of circumstances that requires special consideration. The voltage V1 in figure 4
is a sinusoidal ac voltage, that is
V 1 = V sin (ωt ) (5)

For example, if the three-phase system voltage is nominally 480v, then V1 will have a range of
instantaneous values between zero and the crest value of the 277v. line-to-neutral voltage, or
392v.

There is a minimum voltage required to sustain an arc. The magnitude of this minimum voltage
depends on a variety of factors – whether or not the short circuit is contained in a fashion such
that ionized gas can accumulate in the vicinity of the fault, ambient temperature and humidity,
etc. Generally, however, the minimum voltage is believed to be in the range of 200 – 300v.

As the system line-to-neutral driving voltage varies between zero and its crest value, there will be
periods when the voltage is unable to sustain an arc, and the arc associated with a short circuit
will be extinguished. Obviously, at a subsequent point on the sinusoid of voltage the
instantaneous magnitude may again exceed the threshold voltage required to sustain an arc, but
because the voltage is only slightly greater than the minimum, the arc may not actually restrike in
each successive half-cycle of the driving voltage.

The result is a phenomenon in which arcing ground faults are intermittent. And obviously, if the
arc associated with the short circuit is intermittent, the current that flows in the circuit will also be
intermittent. This has two consequences. First, the effective heating value of the intermittent
short circuit current (also called the rms of the short circuit current) will be lower than the
magnitude of current that one would calculate from known circuit parameters. This means that
protective devices that sense rms current are relatively less effective in detecting the current
associated with arcing ground faults.

Second, if protective devices cannot accurately measure the current associated with arcing ground
faults, they may be unable to detect those faults, thereby allowing the faults to persist for
extended periods. And given the intermittent nature of the arc in an arcing ground fault, the result
is that extensive burning damage can result from the fault.

© 2009 Louie J. Powell page 8 of 20 pages


Power System Neutral Grounding Fundamentals

The classical analysis of the arcing ground fault phenomenon would suggest that arcing ground
faults will only occur on systems where the voltage is 480v or perhaps 600v. At higher system
voltages, the actual instantaneous driving voltage is greater than the minimum voltage required to
sustain an arc for a much greater portion of the voltage half-cycle, with the result that the arc is
more likely to be re-established in every half-cycle of voltage.

Conversely, at lower voltages (for example, on three-phase 208v systems), the actual phase-to-
neutral driving voltage is less than the nominal voltage required to sustain an arc. That leads to
the traditional conclusion that arcing ground faults can’t occur on 208v systems. The fact is that
operators of 208v secondary network systems have experience that conflicts with this theory;
there are many documented instances of intermittent arcing ground faults in 208v network vaults.
The factor that may explain this discrepancy is that network vaults are confined spaces that allow
for accumulation of ionized gas. The effect of this accumulation is to reduce the voltage required
to sustain an arc sufficiently to allow arcing ground faults to occur.

An important consideration that strongly favors solidly or effectively grounded systems is the
ability to support single-phase-to-ground connected loads. It’s not unusual for loads to be
connected single-phase-to ground. In low voltage systems, this is a very practical way to serve
individual loads; 277v lighting is very commonly applied in systems rated 480v three-phase.
Also, the whole point of 208v systems is that the single-phase-to-ground voltage is 120v,
constructing three-phase systems at this voltage allows large numbers of single-phase 120v loads
to be served economically.

In medium- and high-voltage systems, it is less common to see single-phase loads. However, it is
sometimes necessary to supply small aggregations of load from these higher-voltage systems, and
it is economically more appealing to install a single-phase transformer with a single bushing that
is connected phase-to-ground than it is to install a single-phase transformer connected phase-to-
phase and must therefore have two higher-voltage bushings.

When loads are connected on a single-phase basis, the concern is that it is not always possible to
exactly balance the loads across the three phases. Whatever unbalance there may be appears as a
single-phase-to-ground load. Fig 5 is a restatement of the equivalent circuit of fig 4 for the case
of unbalanced single-phase loading.

© 2009 Louie J. Powell page 9 of 20 pages


Power System Neutral Grounding Fundamentals

Fig 5 Equivalent circuit for a solidly grounded system with single-phase-to-ground connected
load

The fact that there is an essentially zero-impedance connection between neutral and ground
means that the only impedances in this circuit are the system inductance, and the impedance of
the load. More importantly, because there is no significant impedance between neutral and
ground means that there is no significant voltage drop between neutral and ground resulting from
the flow of single-phase load current. Therefore, the neutral and ground will have essentially the
same voltage.

If the system were ungrounded, the impedance XC0/3 would appear between neutral and ground.
This would have two consequences. First, because that impedance is relatively large, it would
severely limit the ability of the system to support the single-phase load. Also, to whatever degree
the system did support single-phase loading, the fact that load current would pass through this
impedance would result in a voltage drop between neutral and ground, thereby elevating the
voltages on both the system neutral and on the phase conductor and resulting in a potentially
hazardous condition.

Conclusions: “Solidly” or “Effectively” grounded systems


ƒ Very large ground fault currents
o Limited only by the impedance of the system
o In special cases, may be greater than the three-phase fault current magnitude
o High ground fault current magnitude present potential people hazards
9 Greater arc-flash energy requiring more aggressive personnel protection
9 Earth potential gradient concerns
o Automatic fault detection and clearing is mandatory
ƒ Transient overvoltages associated with creating or clearly ground faults controlled to
within reasonable limits – allows for easier application of surge protective devices
ƒ Potential for hazardous arcing ground faults in some lower voltage applications
ƒ Readily supports single-phase-to-ground connected loading

Case 3: Low resistance connection between N and G

One of the most common options for system neutral grounding in industry involves the
application of a resistor between neutral and ground. Actually, there are two options for resistive
grounding. We will first consider the case where the resistor has a relatively low ohmic rating.
In this context, ‘low’ means that the resistor will have an ohmic value that is much smaller than
XC0/3 and much larger than XL. For example, on a 13.8kV system XC0/3 will have a magnitude in
the range of 1000-2000 ohms, while XL will be a fraction of one ohm. A typical situation would
then be to select a resistor with an ohmic rating of 20 ohms.

Given that

XC 0
>> R >> XL (6)
3

Then, the fault current in this system will be

V1
Ifault = (7)
R

© 2009 Louie J. Powell page 10 of 20 pages


Power System Neutral Grounding Fundamentals

This leads to two important observations:


1. The magnitude of ground fault current will be determined almost exclusively by the
ohmic rating of the resistor, and typically will be limited to a value that is not very
different from load current. For example, in a 13.8kV system, a typical 20 ohm resistor
will limit ground fault current to 400 amperes.
2. The equivalent circuit, shown in fig 6, for a ground fault is therefore a predominantly
resistive circuit. That means that there are no significant switching transients associated
with either applying or clearing ground fault currents.

Fig 6 Equivalent circuit for a system with resistance grounding

The fact that low resistance grounding limits ground fault currents to moderate levels less than
load current has a number of attractive features for industrial applications. Lower ground fault
current means lower arc flash hazards (although it must quickly be pointed out that the arc flash
hazards associated with phase-to-phase and three-phase faults remain a very serious concern).
Also, when less fault current is injected into the ground, the risk of earth potential gradients is
significantly reduced and in fact normally becomes a non-issue. Lower fault current also
normally implies that the burning damage at the point of a fault will be significantly reduced, and
this can be a very important consideration in industrial systems with large populations of medium
voltage motors.

The fact that the circuit is primarily resistive also has advantages. Resistance adds damping to
the equivalent circuit for the fault, so much so that there are essentially no concerns for repetitive
restriking or transient overvoltages associated with ground fault conditions.

The price that has to be paid for these attractive features is that the available ground fault current
is lower than the three-phase fault current, and in fact is not too different from load current. That
raises the concern for fault detection. Fortunately, relaying schemes have been developed that
can very easily discriminate between three-phase balanced load or fault current, and single-phase-
to-ground fault current. The most common feature of these schemes is that they employ
protective relays that measure the residual of the three phase currents rather than the individual
phase currents themselves. Note that conventional fuses tend to be less effective in low-
resistance grounded systems because they respond to currents on an actual per-phase basis.

© 2009 Louie J. Powell page 11 of 20 pages


Power System Neutral Grounding Fundamentals

Also, the fact that a physical resistor is connected between neutral and ground means that under
fault conditions, the system neutral will be elevated from ground potential. As a practical matter,
the neutral-to-ground voltage can be as high as the unfaulted line-to-neutral voltage magnitude.
Resistance grounded systems are clearly NOT effectively grounded systems, and the possible
elevation of neutral voltage means that care has to be taken to insulate and isolate the actual
system neutral.

Also, the elevation of neutral voltage under ground fault conditions means that the voltage on the
unfaulted phases will be elevated to the phase-to-phase voltage magnitude. This results in the
need to apply higher-rated surge voltage protection compared with solidly or effectively grounded
system.

While limiting ground fault current can result in a reduction in fault point damage, this will be the
case only if the fault can be detected and cleared automatically and immediately. In recent years,
industry has come to recognize that faults in the stator winding of generators on low-resistance
grounded systems may actually be exposed to incrementally greater burning damage because
tripping the generator for a stator ground fault does not necessarily immediately clear the fault.5

Finally, the presence of a finite value of resistance between neutral and ground essentially
eliminates the possibility of serving single-phase-to-ground loading.

Conclusions: “Low-Resistance” grounded systems (typically 100-600 amperes)


ƒ Ground fault currents limited to modest values
o Limited almost entirely by the ohmic rating of the resistor
o Typically on the order of load current
o Automatic fault detection and clearing is mandatory
o Requires relaying schemes that measure the residual of the three phase currents
o Traditional fusing may not be a practical form of protection
ƒ No significant concern for transient overvoltages associated with ground faults
ƒ Mandates special considerations in applying surge overvoltage protection
ƒ Does not support single-phase-to-ground connected loading

Case 4: High resistance connection between N and G

If the resistor connected between N and G presents a high magnitude of resistance, some of the
simplifying assumptions made for the low resistance grounding case no longer apply.

As the magnitude of R approaches the scalar magnitude of XC0/3, it is no longer possible to


ignore the effect of distributed capacitance. The term “high resistance grounding” generally
implies the situation in which equation 8 is true.

XCO
R≈ (8)
3

In this situation, the total magnitude of current that will flow to a ground fault on the system will
be

V1
Ig ≈ 2 × (9)
R

© 2009 Louie J. Powell page 12 of 20 pages


Power System Neutral Grounding Fundamentals

Low voltage systems tend to be physically compact primarily because voltage drop is an
impediment to establishing longer circuits. In low voltage systems, it is not uncommon for the
capacitive charging current to be on the order of 1-2 amperes, and low resistance grounding
designed in accordance with equation (8) will result in a total current that is only slightly greater
in magnitude, perhaps 2-3 amperes. With ground faults limited to this extent, it may be possible
to forgo automatic ground fault tripping in favor of allowing the fault to remain while diagnostic
procedures determine the fault location. Since most faults originate on a single-phase-to-ground
basis, the use of high-resistance grounding therefore results in an apparent improvement in
system availability.

The process of locating a fault on a high-resistance grounded system requires that a means be
provided to cause the fault current to pulsate so that it can be distinguished from the natural
distributed charging current in the system. Traditionally, this has been done by applying a
contactor that cyclically shorts a portion of the resistance. An electrician can the trace the
pulsating current using an ordinary clamp-on ammeter.

However, it must be noted that using a clamp-on ammeter to trace a pulsating fault current may
expose the electrician to energized parts and caution must be exercised to verify that the
electrician is equipped with the appropriate personal protective clothing for the arc flash level
available from the system.

Figure 6 is appropriate for the high-resistance grounded case, but in this instance, consideration
has to be given to the fact that the distributed charging capacitance is no longer negligible. The
net impedance between N and G is significantly large. Therefore, when there is a fault on the
system, there will be a significant voltage drop between N and G – the steady-state value of this
voltage will be the nominal line-to-neutral voltage, V1. That means that the system neutral will
be displaced 100% in the event of a ground fault, causing the voltages on the two healthy phases
to have a magnitude above ground equal to the line-to-line system voltage.

The fact that the voltage on the healthy phases is displaced significantly during ground faults may
be a special consideration if the application does not include automatic fault detection and
clearing. The actual insulation capability of power cables is based on the presumption that
voltages will not be significantly unbalanced for extended periods, and if the application
presumes that the fault will be traced manually rather than tripped automatically, it may be
necessary to specify cable with a higher voltage rating.

When the condition of equation (8) is satisfied exactly, the circuit between N and G becomes an
RC circuit with a time constant of one electrical radian, and that fact is very significant. Earlier,
in the discussion of the ungrounded neutral (case 1), it was observed that it was possible for
voltage to gradually build up across the equivalent system charging capacitance, XC0/3, resulting
in devastatingly high phase-to-neutral voltages. With a time constant of one electrical radian, it is
not possible for this voltage escalation effect to take place. Whatever voltage is trapped on XC0/3
during a fault will decay rapidly preventing the step-wise escalation of voltage described earlier
as repetitive restriking. That consideration is another of the strong advantages of high-resistance
grounding.

Historically, high-resistance grounding was originally intended for low-voltage applications. It


has also been widely used in some higher voltage applications, but there are some serious
constraints that must be observed. The key point is the objective for using high-resistance
grounding. There are basically two scenarios here:

© 2009 Louie J. Powell page 13 of 20 pages


Power System Neutral Grounding Fundamentals

• If the objective for high-resistance grounding is simply to limit the available ground
fault current, thereby limiting burning damage and earth potential gradients, then the
system can be applied along with equipment that automatically detects faults and
initiates automatic tripping. High resistance grounding is very commonly applied at
the neutrals of generators connected to the grid through dedicated ‘unit step-up’
transformers. In these applications, even at unit voltages of 20kV and greater, the total
natural distributed charging current rarely exceeds 5-7 amperes, and high-resistance
grounding systems are typically designed to limit ground fault current to 10 amperes.

Another common situation for high-resistance grounding with automatic tripping is in


mining applications. There are special safety rules in the mining industry that limit the
maximum earth potential gradients, and to achieve that level of personnel protection, its
common to see high-resistance grounding used to limit ground fault currents to 25
amperes.

• But if the objective for high-resistance grounding is to provide a traceable ground fault
so that a diagnostic procedure can replace automatic tripping, then special
consideration has to be given to the maximum magnitude of ground fault current in
higher-voltage applications. Empirical evidence suggests that if the magnitude of
ground fault current exceeds 10 amperes, the amount of burning that will take place at
the fault point will increase and eventually lead to escalation into a multi-phase fault.
The natural charging current present at system voltages greater than 5kV is often
greater than 10 A. For that reason, when high-resistance grounding is applied at system
voltages in excess of 5kV, it is usually recommended that the system be designed to
automatically detect and trip for ground faults.

In low voltage applications, the circuit shown in fig 6 applies literally. In higher voltage
applications, however, it is common to see a small transformer connected between N and G, with
a resistor across the secondary of the transformer. In this arrangement, the effective magnitude of
resistance in the circuit is the actual ohmic rating of the resistor multiplied by the square of the
transformer turns ratio. This typically results in a lower cost, more compact installation than
would be the case if the resistor were fully rated for the required resistance. The transformer
would normally have a primary voltage rating equal to (or perhaps greater than) the line-to-
neutral voltage rating of the system, and would require a thermal (or kVA) rating sufficient to
withstand the loading associated by a ground fault for time that the fault will be allowed to persist
on the system.

Conclusions: “High-Resistance” grounded systems


ƒ Ground fault currents limited to very low values
o Fault current limited to slightly greater than the natural system distributed
charging current
o Fundamental design criterion:
XCO
R≈
3
• Traceable fault current option
o Uses a contactor to cyclically short a portion of the resistor causing the fault
current to pulsate
o Fault location is possible using a hand-held clamp-on ammeter

© 2009 Louie J. Powell page 14 of 20 pages


Power System Neutral Grounding Fundamentals

o Eliminates the need for automatic ground fault tripping, resulting in an apparent
increase in system availability
o Electricians MUST be equipped with appropriate personal protection gear
o May require impose special considerations in specifying power cable insulation
o Should not be applied in systems rated above 5kV
• In other applications
o Mining, voltages greater than 5kV, and unit-connected generators
o Objective is only to limit burning damage and potential gradients
o Requires automatic tripping
• No significant concern for transient overvoltages if design criterion is met
ƒ Sustained overvoltage on healthy phases during ground faults mandates special
considerations in applying surge overvoltage protection
ƒ Does not support single-phase-to-ground connected loading

Case 5: Low (inductive) reactance connection between N and G

Yet another possibility is to connect an inductance between neutral (N) and ground (G) as shown
in fig 7. As in the case of resistance, this comes in two ‘flavors’ – low inductance and high
inductance.

Fig 7 – Equivalent circuit for a low-inductance grounded system

The first observation that can be made about low reactance grounding is that while the elements
between neutral and ground form a parallel L-C circuit, the inductive reactance is many orders of
magnitude smaller than the capacitive reactance, XC0/3. Therefore, the capacitance can often be
neglected.

There is resistance in this circuit. The typical X/R ratio of an inductor falls in the range of 60-
100, while the X/R ratio of typical system inductive reactances is 5 to 15 (low voltage systems
tend to have a lower X/R than do higher voltage systems). And while the resistance is significant
and must be considered in calculating the actual magnitude of short circuit current for the purpose
of evaluating the duty on circuit breakers6, it is possible to ignore the resistive component of these
impedances while examining the fundamental concepts involved in the application.

© 2009 Louie J. Powell page 15 of 20 pages


Power System Neutral Grounding Fundamentals

Because the circuit has the same fundamental R-L characteristics of the basic power system,
therefore, the practice of low reactance grounding does not introduce any new concerns with
respect to switching and transients.

While the equivalent circuit of fig 1 and its subsequent descendents is literally correct, this
discussion has intentionally overlooked the question of whether these equivalent circuits contain
all of the information needed to determine the actual magnitude of ground fault current for the
applications they represent. The fact is that there are some additional considerations that enter
into the determination of ground fault magnitudes, especially if the system neutral is solidly or
effectively grounded (Case 2, fig. 5). When those factors are considered, one can conclude that
the magnitude of ground fault current can sometimes exceed the available magnitude of current
that would flow in the same system to a balanced, three-phase fault7. Those special
circumstances include:
• Applications in the immediate vicinity of the solidly-grounded wye connection of
delta-wye transformers (or three-phase banks of single-phase transformers).
• Applications at the terminals of generators whose neutrals are solidly grounded
The most common situation for applying low reactance grounding is in one of these two
situations where it is desired to ‘fine-tune’ the available magnitude of ground fault current to a
slightly lower value.

Consider the multiple-transformer substation situation. While it is a relatively unusual situation,


it is possible for the single-phase-to-ground fault level to exceed the interrupting ratings of circuit
breakers even though those breakers have sufficient capability to switch the available three-phase
fault level. This situation tends to come about when there are multiple delta--grounded-wye
transformers connected together at their wye-connected terminals. In such instances, one solution
is to install neutral grounding reactors (low inductance grounding) in between the neutral
terminals of the transformers and ground. Whether the design limits the available single-line-to-
ground fault level to equal the three-phase fault level, or whether an arbitrary ground fault level is
merely assumed as a design objective is a decision that the system engineer must make based on
both technical and economic considerations.

In the generator instance, any time the neutral of a generator is solidly grounded, the available
ground fault level will exceed the three-phase fault level. Obviously, this is not a problem for
generators that are intended for installation in systems where the neutral is solidly grounded
(especially, low voltage generators), but it can be a problem with larger (higher voltage)
generators where NEMA standards do not require that the machines be mechanically braced for
unbalanced fault stresses that are greater than the stresses associated with balanced three-phase
faults. Again, the solution is to install low-inductance grounding at the generator neutral.
Usually, the practice is to design the low-reactance grounding installation to limit the single-line-
to-ground fault to equal the three-phase fault, but it is necessary for the system engineer go also
consider other technical and economic factors before finalizing the design.

The criteria used in designing a low-reactance grounding application depend very much on the
objectives that the system is expected to support. Design is almost always a matter of
compromising between competing objectives. For example, it is possible to design a low-
reactance grounding application that will deliver “effective grounding” (as defined by equations
[3] and [4]. It is also possible to design a low-reactance grounding application that supports
single-phase-to-ground loading (ie, the effective ‘low reactance’ inserted between N and G is
small enough that the voltage drop between N and G due to unbalanced single-phase loading is
acceptable. On the other hand, such designs may not limit the available single-line-to-ground
fault current sufficiently to address a circuit breaker rating problem.

© 2009 Louie J. Powell page 16 of 20 pages


Power System Neutral Grounding Fundamentals

And commonly, another limitation is the practicality and economics of the design. The cost of
neutral grounding reactor increases as they are designed to have lower inductance, and there is a
threshold below which it is not practical to build an inductor. As a consequence, it may not be
economically practical to design a solution that requires only a very small incremental neutral
inductance.

Low-reactance grounding is one of the least common forms of system neutral grounding. The
need to address multiple competing design objectives, and the amount of work involved in
evaluating those options, are such that it is typically reserved for one of the two special cases
cited.

Conclusions: “Low Reactance” grounded systems


• Ground fault currents limited to meet a system design criterion
o Limit ground fault current to an objective magnitude
o Limit ground fault current to no greater than three-phase fault magnitude
• May or may not achieve ‘effective grounding’ depending on design objectives
• May or may not support single-phase-to-ground connected loading depending on design
objectives
• A relatively uncommon practice reserved for special applications
• Requires extensive system engineering

Case 6: High (inductive) reactance connection between N and G

As the name implies, high reactance grounding involves installing a high magnitude of inductive
reactance between neutral (N) and ground (G). Because the inductive reactance is high, it is not
possible to ignore the fact that the presence of this inductance makes the connection between N
and G a parallel L-C circuit. In fact, it is the nature of that L-C circuit that gives this option its
most attractive features.

The design criterion for a high-inductive grounding system is given in equation 10

XCO
jXG = [10]
3
When this criterion is met, the fundamental frequency impedance between N and G is infinite –
an open circuit. As a result, in steady state, high-reactance ground acts just like the ungrounded
case, case 1. Specifically, there can be no ground fault current. In turn, that means that there is
no fault-point damage, no mechanical distress on current-carrying conductors, no thermal
distress, no earth potential gradients, and no arc flash.

But there is still an electrical connection between N and G that presents impedances at
frequencies other than the fundamental frequency. Therefore, those undesirable consequences of
an ungrounded system that are mainly related to its performance with respect to non-fundamental
frequency phenomena such as transient response and overvoltages, also do not appear.

Historically, this form of grounding, also known as “resonant neutral grounding” or “Peterson
Coil Grounding” (in honor of its inventor), was used to some degree in high-voltage transmission
applications in North America up through the middle of the 20th century. At that point, however,
its major drawback began to become an obstacle. Unlike high-resistance grounding, where it is
necessary that the resistance only approximate the distributed charging reactance, in order for

© 2009 Louie J. Powell page 17 of 20 pages


Power System Neutral Grounding Fundamentals

resonant grounding to work, the design criterion for resonant grounding must be met exactly.
That means that each installation must be ‘tuned’ to the distributed capacitance that it encounters
in the system where it is installed. As transmission grids started expanding in the 1940’s and
50’s, it became necessary to provide a means of tuning the reactor to compensate for switching
events on the system that changed the charging capacitance. Eventually, the resonant neutral
grounding applications on the transmission system in North America were retired.

However, there were transmission-level resonant neutral grounding applications in service in


other parts of the world well into the 1970’s, particularly in Asia.

Another area where resonant neutral grounding was applied was at the neutrals of generators. In
theory, these applications should have been ideal because, with the generator connected to the
system through a dedicated generator step-up transformer, the only XC0/3 that would need to be
considered was the capacitance of the generator, generator leads, and the delta-connected winding
of the transformer. Practically, that means that the XC0/3 would essentially be the surge
capacitors applied directly at the terminals of the generator. And the fact that resonant grounding
would eliminate the risk of fault-point burning was a very attractive feature.

On the other hand, experience with high-resistance grounding proved that it was just as effective
in managing fault point burning, and the overall cost of the resistive approach was lower than the
cost of a resonant solution. Even so, there were a number of power plants, mainly in the New
England area, that were built with resonant grounding of their generators8.

Today, however, resonant grounding is mainly an academic topic in North America. That is not
to say that the practice has been abandoned. In fact, it is in very widespread use in medium-
voltage utility distribution applications in Europe and the UK.

Conclusions: “High Reactance” grounded systems


• Ground fault currents limited to essentially zero
o Design criterion
XCO
jXG =
3
o Design criterion must be met exactly – requires retuning if switching changes
distributed capacitance of the system
• No unusual concerns for transient overvoltages
• Non-effective grounding – requires special considerations when applying surge
protection
• Does not support single-phase-to-ground connected loading
• No known installations in North America, but in wide-spread use in Europe and UK
• Also known as “Peterson Coil” or “resonant neutral’ grounding

Applications Summary

It is obviously impossible to declare that there is a single set of universally-correct answers.


However, it is possible to summarize what appears to be the general application trends in system
neutral grounding.

© 2009 Louie J. Powell page 18 of 20 pages


Power System Neutral Grounding Fundamentals

Grounding Practice Usual Applications


Ungrounded neutral • Generally not recommended on new systems
• May exist on legacy systems although a retrofit is usually
recommended.
• Marine systems may be a special case
Solidly grounded neutral • Preferred practice for high voltage transmission systems
(“effective grounding”) • Preferred practice on medium-voltage utility distribution
systems in North America
• Commonly applied on low voltage systems serving single-phase-
neutral loading
• Not recommended for medium-voltage distribution in industrial
workplaces
Low resistance grounding • Preferred practice for medium-voltage distribution in industrial
workplaces
High-resistance grounding • Preferred practice for unit-connected generator applications
(with automatic tripping)
• Commonly applied in continuous process industrial applications,
5kV and below in conjunction with traceable fault technology
• May be used at higher voltages with automatic fault detection
and tripping
• Cost-effective retrofit for legacy ungrounded systems
Low reactance grounding • Solution for managing high ground fault currents in substations
• Solution for generator grounding to support single-phase loading
• Requires extensive application engineering
High reactance grounding • Rarely seen in North America today
• Commonly applied in medium voltage utility distribution
applications in Europe and UK
Other practices – including These approaches have been used as retrofit solution for legacy
‘corner of the delta’ and ‘mid- ungrounded systems. In general, these are compromise solutions and
point’ grounding are not currently recommended practices.

Bonding and Grounding

This course has focused on the issues and practices involved in determining how the neutral of
the electrical system should be connected to earth or ground. The primary objective of that focus
is managing the magnitude of current that will be injected into ground as a result of a fault to
ground on the electrical system.

There is a closely related set of concerns associated with the design of ground grids, and the
practices of bonding metallic structures. These concerns address management of the potential
gradients that will arise as a result of that fault current being injected into ground. The primary
criteria in establishing bonding practices is to limit the voltage to which a person can be exposed,
either as a consequence of touching a metallic structure (where the voltage is the difference
between the potential where the person is standing and the potential on the structure) or as a
consequence of walking through a facility (where the voltage is the difference of potential across
the length of the person’s stride).

Bonding and ground grid design is critical to personnel safety and is a specialized field in and of
itself. It is mentioned here in order to emphasize that while establishment of proper system

© 2009 Louie J. Powell page 19 of 20 pages


Power System Neutral Grounding Fundamentals

neutral grounding is important, it is not the complete answer and must go hand-in-hand with
competent design of ground grids and bonding practices.
1
Dunki-Jacobs, JR; Shields, FJ and St Pierre, CR; The Industrial Power System Grounding Design
Handbook, http://groundingdesignbook.com/index.html.
2
IEEE Std 80™-1986, IEEE Guide for Safety in AC Substation Grounding
3
IEEE Std 142™-1991, IEEE Recommended Practice for Grounding of Industrial and Commercial
Power Systems (The IEEE “Green Book”).
4
Peterson, Harold A., Transients in Power Systems, John Wiley & Sons, New York, 1951; reprinted
1966, Dover Publishing, New York.
5
Powell, Louie J. “The impact of system grounding practices on generator fault damage,” IEEE
Transactions on Industry Applications, vol. IA-34, Sept./Oct. 1998, pp. 923-927.
6
IEEE Std 551-2006™-, IEEE Recommended Practice for Calculating Short-Circuit Currents in
Industrial and Commercial Power Systems (The IEEE “Violet Book”)
7
Students interested in expanding their understanding of these considerations are encouraged to take the
PDHengineer course on Symmetrical Components, E-4002
8
The reason these installations tended to be in New England was that the main proponent of resonant
grounding of generators was the legendary Prof Eric T. B. Gross, initially at Cornell and later at Rensselaer.
Many of his students spent their careers close to home in the New England area, and served as his disciples
in advocating for this solution.

© 2009 Louie J. Powell page 20 of 20 pages

You might also like