GER-3793

GROUND DISTANCE RELAYING:

PROBLEMS AND PRINCIPLES

G. E. Alexander
J. G. Andrichak

General Electric Company

Meter and Control Business Department
Malvern, PA

Presented to the
Nineteenth Annual
Western Protective Relay Conference
Spokane, Washington
October 20, 1991
 GROUND DISTANCE RELAYING:
PROBLEMS AND PRINCIPLES

G. E. Alexander
J. G. Andrichak
General Electric Company
Malvern, PA

INTRODUCTION
The single line to ground fault is the most common type of fault
that occurs on transmission lines, and ground distance functions
are often used for the detection of these faults. However,
ground distance functions do have acknowledged limitations and
special consideration should be given to the design and
application of the functions. This paper will discuss the
operating principles of ground distance functions and some of the
limitations, or problems, that can be encountered in their
application.
Ground distance functions measure an apparent impedance to the
fault based on the ratio of relay voltage to relay current. A
variety of ground distance relay characteristics have been used:
reactance, mho, and quadrilateral for example.
The simplest ground distance functions use only a single phase
current and a single phase voltage. As the design of ground
distance functions has improved, other currents and voltages
have been added to the relay operating quantities. This paper
will discuss distance functions of the "Phase angle comparator"
design as discussed in the next section. A "classical" mho unit
has two inputs: an operating signal (IZ - V), and a polarizing
signal (VPOL). Some of the newer relay designs employ three or
more inputs to the coincidence logic. These additional inputs
greatly improve the performance of the relay, but they may
require a more complex test setup to insure that all of the
inputs have the proper phase relationship.

DEVELOPMENT OF THE MRO CHARACTERISTIC

INPUT SIGNALS
Mho distance type characteristics may be developed by measuring
the phase angle between two or more voltage signals in the relay.

-l-
The logic used to measure the phase angle is called a Phase Angle
Comparator. For the simplest two input comparators, the voltage
signals in the relay are referred to as the operating signal and
the polarizing signal. These signals are derived from the system
voltage and the system current supplied to the relay.
In an analog relay, the relay current is converted to a voltage
in the relay, typically by a transactor. A transactor is an iron
core reactor with an air gap. The transactor produces a voltage
proportional to the input current. The transfer impedance of the
transactor is defined as ZR, and is used to determine the reach
and angle of maximum reach of the mho characteristic. Other
techniques may be used in digital relays: for example, the
I·ZR - V and V signals may be developed using a digital Fourier
filter. The Fourier produces a phasor output: that is, the
I·ZR - V and the V have both magnitude and phase angle. The
phase angle comparator is a software module that checks if the
relative phase angles of the operating and polarizing signals are
within the tripping range, based on the desired characteristic.
Mho distance characteristics are typically plotted as circles on
an R-X impedance diagram as shown in Figure l. However, because
the m h o characteristic is actually determined by the angle
between voltage phasors, it is convenient to plot the

TYPICAL MHO CHARACTERISTIC TYPICAL MHO CHARACTERISTIC

R-X IMPEDANCE DIAGRAM IR-IX VOLTAGE DIAGRAM

Figure 1 Figure 2

characteristics on a voltage diagram to describe how they are

derived. The voltage diagram may be obtained from the R-X
diagram by multiplying all the points by the fault current in the
relay as shown in Figure 2. Since the fault current will change
as the system conditions and fault location change, the voltage
diagram will contract or expand for different fault currents. The

-2-
voltage phasors, however, will have the same phase angle and
magnitude relationships as the impedance vectors on the R-X
diagram.
Consider the simple system shown in the one line diagram of
Figure 3. Let ZR be the reach of the mho function, ZF be the
impedance from the relay location to the fault, and V and I be
the current and voltage supplied to the relay. For the fault on
Figure 3, the voltage V supplied to the relay will be I·ZF. The
current I supplied to the relay is transformed into a voltage
I·ZR by the action of the transactor. The actual current and
voltage used by the mho function depend upon the individual
relay.

Figure 3
THE IR-IX DIAGRAM
The "Classical" mho characteristic is shown as a circle on the
IR-IX diagram, passing through the origin as shown in Figure 4.
The diameter of the circle which passes through the origin is
referred to as the "angle of maximum reach", or in
Electra-Mechanical relay terms, the "angle of maximum torque".
The angle of maximum reach of the circle, I·ZR is determined by
the replica impedance of the function, ZR. If the voltage phasor
V shown on Figure 4 is the voltage supplied to the relay under
fault conditions, then the position of V is along the angle of
the impedance to the fault, ZF. The magnitude of V will
therefore depend on the fault location. For a fault at the
balance point of the relay, the V will plot on the mho circle as
shown in Figure 4.
From geometry it is known that if one side of a triangle
inscribed in a circle is the diameter of the circle, then the
angle opposite the diameter is 90 degrees. Since I·ZR is the
diameter of the circle, both angles A and B in Figure 4 are equal
to 90 degrees. Two sides of the triangle are formed by the
phasors V and I·ZR, the third side is the phasor difference of
these signals, I·ZR - V, which is known as the operating signal
of the mho function. The locus of all points for which the
phasors V and I·ZR - V are 90 degrees apart, is a circle that
passes through the origin. The phasor V is known as the
polarizing signal. For the classical mho function, also known as
 BALANCE POINT

Figure 4
a self-polarized mho, this voltage is the fault voltage used to
form I·ZR - V. Sine wave signals representing V and I·ZR - V are
shown in Figure 5 corresponding to the phasors of Figure 4. For
this condition, I·ZR - V leads V by 90 degrees.

MH0 CHARACTERISTIC BALANCE POINT

IZR-V LEADS VBY 90 DEGREES

Figure 5
If the fault is moved closer to the relay, that is, inside its
zone of protection, the magnitude of V will decrease relative to
the magnitude of the I·ZR as shown in Figure 6. For this
condition, the angle A is greater than 90 degrees while the angle
B is less than 90 degrees. For this condition, as shown in
Figure 7, I·ZR - V leads V by less than 90 degrees. As the fault
is moved farther from the relay, that is, outside its zone of
protection, the magnitude of V will increase relative to the
magnitude of the I·ZR as shown in Figure 8. For this condition,
the angle A is less than 90 degrees while the angle B is greater
than 90 degrees. For this condition, as shown in Figure 9, I·ZR
- V leads V by more than 90 degrees.
The fault is within the m h o function's zone of protection if
I·ZR - V leads or lags V by 90 degrees or less. This is
-4-
depicted in Figure 10. Thus a mho characteristic can be created
if the phase angle between these two signals can be accurately
measured.

INTERNAL FAULT
RR-V LEADS V BY LESS MAN 90 DEGREES

INTERNAL FAULT

Figure 6 Figure 7

EXTERNAL FAULT
IZR-V LEADS V BY MORE THAN SO DEGREES

EXTERNAL FAULT

Figure 8 Figure 9

902
_______________ k 900
___________------

RANGE OF OPERATING ANGLES

Figure 10
IZ-V
PHASE ANGLE COMPARATOR LOGIC
Phase Angle Comparator Logic is used to measure the phase angle
between I·ZR - V and V in the implementation of the mho function.
The phase angle comparison can be an analog circuit or software.
A simple phase angle comparator circuit is shown in Figure 11.
This circuit is typical of the comparator used in all analog
solid state GE m h o distance relays. This type of circuit
measures the coincidence between the operating and polarizing
waveforms. Coincidence occurs when both signals have the same
instantaneous polarity, as determined via the AND functions of
Figure 11. The coincidence blocks are directed to the input of
the characteristic timer shown in Figure 11; the characteristic
timer produces an output if the duration of the coincidence is
longer than its pickup setting. For a circular m h o
characteristic the pickup of the timer is set for 90 electrical
degrees, or 4.16 milliseconds for a 60 hertz system.

PHASE ANGLE COMPARATOR

t 2 INPUT I

Figure 11

The shape of the mho characteristic on the R-X diagram may be

modified by changing the pickup setting of the characteristic
timer. Figure 12 shows the effects of increasing and decreasing
the setting from 90 degrees. The lens characteristic is created
by increasing the pickup setting to a value greater than 90
degrees: the expanded mho is the result of a pickup setting less
than 90 degrees.

PICKUP < 90 DEGREES PICWC > 90 DEGREES

EFFECT OF CHARACTERISTIC TIMER SETTING

Figure 12
-6-
The Phase Angle Comparator may be used to develop relay
characteristics other than the mho characteristic by changing the
input signals. Table I shows a selection of various ground relay
characteristics that can be obtained from a simple two input
Phase Angle Comparator.
TABLE I
CHARACTERISTIC OPERATING SIGNAL POLARIZING SIGNAL
MHO IZ-v V
REACTANCE IZ-v IZ
DIRECTIONAL IZ V
The design of the phase angle comparator ground mho distance
function can be extended to include more input signals as shown
in Figure 13. In this case, the coincidence between all of the
input signals must be at least equal to the characteristic timer
setting. This implementation includes several characteristics:
the IZ-V and V form a mho characteristic, the IZ-V and 12Z form a
negative sequence reactance characteristic, and the IZ-V and the
IOZ form a zero sequence reactance characteristic

FOUR INPUT
PHASE ANGLE COMPARATOR

Figure 13

OPERATING QUANTITIES
Ground distance relays are typically designed to be set based on
the positive sequence impedance to the fault for single line to
ground faults. The impedance seen by the relay is a combination
of the positive, negative, and zero sequence impedances of the
system. The positive and negative sequence impedances are
generally equal to each other in magnitude and angle while the
zero sequence impedance may have both a different magnitude and
angle. Because of this, ground distance relays generally employ
some form of "zero sequence current, compensation" to adjust the
measured impedance to reflect the actual system impedances.

-7-
The operating quantity on GE relays is typically compensated as
shown:
(IA - IO)·ZRl + KO·IO·ZRO - VAG (1)
Where:
IA is the faulted phase current
IO is the zero sequence current
VAG is the faulted phase voltage
ZRl is the relay positive sequence replica impedance
ZRO is the relay zero sequence replica impedance
KO is the zero sequence current compensation factor
Since IA is equal to IA1 + IA2 + IAO, where IAl, IA2, and IA0 are
the sequence current components referenced to IA, the operating
quantity may also be expressed as:
(IA1 + IA2)·ZRl + KO·IO·ZRO - VAG (2)
The sequence network connections for the radial single line to
ground fault of Figure 14 are shown in Figure 15. In the design
of most GE relays, the magnitude of ZRl and ZRO are equal by
design. The current compensation factor, KO, is a scalar set
equal to the ratio of the magnitude of the zero sequence
impedance of the protected line to the magnitude of the positive
sequence impedance of the protected line.
Table II shows the currents and voltages for a phase A to ground
fault at the end of the protected line of Figure 15. Assume that
the reach of the ground relay is set equal to the line length, 6
ohms. The apparent impedance seen by a compensated relay can be
calculated by:
VAG
(3)
IA1 + IA2 + KO .I0
Note that the KO used in equation 3 is a phasor rather than the
scalar used in equations 1 and 2. This is to account for the
angular difference in the positive and zero sequence impedances.
TABLE II
11 = 1.95 A l-79.8'
12 = 1.95 A j-79.8*
IO = 1.95 A f-79.8'
VA = 58.3 v /-0.8'
IA = 5.85 A /-?%8’
KO = Z0L/ZlL = 3 f&Q'

-8-
 RELAY FAULT
LOCATION LOCATION RF”,Go_O

- f

El=67 AT 0 D E G

85 deg Z1=6.0 at 85 deg

Z0=1.5 at 85 deg Z0=18.0 at 75 deg

Figure 14

RELAY FAULT
LOCATION LOCATION

El=67 AT 0 DEG
V1
I I1
POSITIVE
Tvy*n f

ZS1 ZL1
1.5 at 8 5 6.0 at 8 5

NEGATIVE v2 I I2
I *’ - I

ZERO
_
zso
vo
I_ IO

ZLO
1.5 at 85 18.0 at 75

Figure 15

-9-
The apparent impedance given by Equation 20 is equal to 6 ohms at
85 degrees. If the impedance had been calculated by simply
dividing the phase A voltage by the phase A current, the result
would have been 10 ohms at 79 degrees, which of course is not the
correct positive sequence impedance to the fault.

POLARIZING QUANTITIES
A mho distance relay function may be developed using a variety of
voltages as the polarizing signal. The Classical, or
self-polarized, mho function discussed thus far uses the faulted
phase voltage as the polarizing signal. That is, a phase A
ground function would use VA, and a phase AB phase function would
use VAB. The resulting mho characteristic, when plotted on the
R-X diagram, remains invariant under changing system and fault
conditions. In 1949, A. R. van C. Warrington published his
classic paper "Graphical Method for Estimating the Performance of
Distance Relays During Faults and Power Swings" [10]. The
methods presented in this paper are widely used to visualize the
performance of distance relays on the R-X diagram. However,
Warrington's paper contains the following sentence:
“It must be remembered, however, that
distance relays polarized by potential other
than the one used for the restraint windings
have a family of circles for different fault
conditions and hence their characteristics
are more difficult to show on an impedance
diagram."

Most modern relay designs use part or all of the unfaulted phase
voltages in the polarizing signal. This is often referred to as
"cross polarization@', orr "healthy phase polarization". Thus the
actual mho characteristics of the relay do not have an invariant
plot on the R-X impedance diagram: rather, they possess an
infinite number of characteristic plots which vary based on power
system conditions and fault type.
POLARIZING SIGNAL REQIUREMENTS
The primary requirement for the polarizing voltage is that the
pre-fault phase angle of the polarizing signal be the same as
that of the restraint voltage used in the operate signal. It is
also desirable that, if possible, the polarizing voltage
magnitude during a fault will be sufficient for the mho function
to operate steady state during a fault in which the faulted phase
voltage is zero. Consideration should also be given to the
effect on the distance relay if a fuse in the secondary of the
potential device fails. When a ground distance mho function is
used in a single phase tripping scheme there is an additional
requirement: the polarizing voltage on the sound phases should
-10-
not be affected by the lack of voltage, or by spurious voltage,
on the open phase during the open pole period following a single
phase trip.
Before attempting to determine the "ideal" polarizing voltage for
a ground distance unit, performance criteria on which to base the
decision must be selected. Depending upon the intended
application of the relay, the following points may be of
importance:
Operation for zero voltage ground faults
L'. Directional integrity
Fault resistance/arc impedance accommodation
z: Correct phase selection in single phase tripping
schemes
e. Proper operation with one phase open in single pole
tripping schemes
MEMORY ACTION
In the "Classical" mho unit, the polarizing voltage is the same
as the restraint voltage used in the operate signal: the faulted
phase voltage. This presents a problem when a fault is applied
to the line at the relay location, since the faulted phase
voltage goes to zero magnitude (neglecting arc/fault impedance).
In order to overcome this situation, "memory" circuits were
added. The memory circuit is designed to provide a "remembered"
voltage, based on the prefault voltage, to the relay for a long
enough time to allow the function to operate for an internal
fault. This circuit would typically be a tuned circuit in
electro-mechanical relays, an active bandpass filter in static
relays, and some form of digital memory in a microprocessor based
relay. The time duration of the memory action is determined by
the design of the particular relay. Relay designs that use
"cross polarization", or "healthy phase polarization", may not
require memory action for single line to ground faults since the
polarizing voltage will not go to zero. However, for a three
phase fault, all voltage to the relay is removed and thus memory
action is still required for three phase faults if the ground
distance functions must operate for three phase faults.
A side effect of the memory action is the introduction of a time
varying response in the mho unit: that is, the response, or
"characteristic", initially following a fault is not the same as
the "steady state" response, or characteristic. This will be
discussed in more detail in a subsequent section.
CHARACTERISTIC TERMS
As was mentioned earlier, mho distance functions typically have
more than one characteristic during fault conditions when viewed
on an R-X diagram. GE uses several terms to describe these
characteristic as explained below.
-ll-
 Dynamic Characteristic: The dynamic characteristic is the
initial characteristic of a time varying impedance
characteristic. The variation with time is due to the
memory action of the relay.
Variable Characteristic: The variable characteristic of a
mho function is due to the use of some form of healthy phase
polarization. The characteristic varies as a function of
the relay design, the power system, and the fault type. The
variable characteristic does not vary with time.
Steady State Characteristic: The steady state
characteristic is the Classical mho characteristic that
passes through the origin on the R-X diagram. this is the
mho characteristic that is typically plotted during routine
relay testing. Depending upon the relay design, this
characteristic may, or may not, exist during actual fault
conditions.
All of these characteristics may exist in the same relay design.
The mho function may have a variable characteristic, a dynamic
characteristic, and a steady state characteristic, as well. Both
the dynamic and variable mho characteristics include more area on
the R-X diagram than the Classical Mho circle. The dynamic
characteristic and the steady state characteristic for a faulted
phase polarized ground unit are shown in Figure 16. It should be
noted here that although the characteristic includes an area
below the R axis, the unit is still directional. The mho
function will only operate for faults below the R axis if they
are capacitive faults in front of the relay (that is, in the
"trip" direction) .

STEADY STATE
CHARACTERISTIC

Figure 16

-12-
CHOICE OF POLARIZING SIGNALS
There has been a wider variety of polarizing voltages used in
ground distance functions than in the phase distance functions.
Although the types of polarizing signals discussed in this paper
will be limited, it must be recognized that other voltages may be
used. Typical signals for a phase A function are:
(VAG)M : faulted phase voltage with memory
(VBC L 90°)M: quadrature voltage: phase BC voltage
rotated 90 degrees with memory.

(VAM )M : median, phase A to the midpoint of VBC

with memory.
(VAGl )M : positive sequence voltage referenced to
phase A with memory.

vc VB
VBC

FAULTED PHASE QUADRATURE

vClvB
]VA-D.S-+VC~C~-1~~A+~~vB+~2-
1

VB
VBC

MEDIAN POSITIVE SEQUENCE

Figure 17
The size of the variable characteristic is a function of the
choice of polarizing voltage, the power system configuration, and
the fault type. Consider the simple radial system of Figure 18.
Figure 19 is a plot of the dynamic and variable characteristics
of four ground distance units with different polarizing voltages
for the same single line to ground fault. The dynamic
characteristic is the same for each of these relays, and is the
same as the variable characteristic for quadrature polarized
function. The greatest fault resistance coverage is provided by
the quadrature polarized unit: the least by the faulted phase
polarized unit. The equivalent source impedances used to
determine the variable characteristics for these conditions are :

POLARIZING VOLTAGE EQUIVALENT SOURCE

IMPEDANCE

Faulted Phase' : 0
Median : (1/3)·Z0S
Positive Sequence : (1/3)·(ZlS + ZOS)
Quadrature : (2/3)·ZlS + (1/3)·ZOS

RELAY FAULT
LOCATION LOCATION

/ r POSITIVE SEQUENCE

Figure 18
Figure 19

ZOL
RELAY
LOCATION

Zls’ = Z1S + Z1L

zos’ = zos+ ZOL
R
Figure 21
Figure 20

As was stated previously, these variable and dynamic

characteristics remain directional even though the plot extends
below the R axis. In fact, the relays possess a totally
different characteristic lot for faults in the reverse
direction. Figure 20 shows t e plot of the quadrature polarized
mho function for a single line to ground fault behind the relay
location, as shown in Figure 21.

-14-
 EFFECTS OF MUTUAL INDUCTION ON GROUND DISTANCE RELAYS

MUTUAL INDUCTION

When two or more lines are in parallel on the same right of way,
fault current in one line may induce voltages in the parallel
line or lines. A typical configuration of 345 KV lines on a
double circuit tower, with close coupling, is shown in Figure 22.
Typical self and mutual impedances are shown in Table III.

I4
0 0
0* 00 00 QA’ -
TABLE III
16
0% 00 00 40’ - Z1 0.06 + j 0.59 ohms/mile
=
16 Z0 = 0.55 +j 1.96 ohms/mile
6C 00 oogc’ - ZlM = 0.01 + j 0.05 ohms/mile
I 28’ I ZOM = 0.49 + j 1.30 ohms/mile

Zl : Positive sequence self

88’
impedance
Z0 : Zero sequence self
impedance
ZlM : Positive sequence mutual
- impedance
ZOM : Zero sequence mutual
impedance

The zero sequence mutual impedance is about two-thirds of the

zero sequence self impedance. By contrast the positive and
negative sequence mutual impedances are on the order of
one-twelfth of the self impedances. Consequently, the positive
and negative sequence mutual impedances can usually be ignored in
the application of distance relays. The zero sequence mutual
impedance, on the other hand, is significant and its effect on
ground distance relays must be considered.
EFFECT ON DISTANCE MEASUREMENT
Depending on the power system configuration in the vicinity of
the relay location, the presence of mutual induction will cause a
ground distance relay to see an apparent impedance that is larger
than or smaller than the actual impedance to the fault. A
typical system involving mutual coupled lines is shown in Figure
23. Two cases will be cited in which the mutual coupling between

-15-
lines 1 and 2 will cause a ground distance relay on line 1 at
breaker A to see a smaller impedance:
Case 1: Tie breaker F open, a single line to ground fault at
bus Y, and the zero sequence current in line 2, from D
to E, is a large percentage of the zero sequence
current in line 1.
Case 2: Breakers C and D are open, a single line to ground
fault at bus Y, and grounding chains are used at both
ends of the line to ground all three phase conductors.

BUS Y BUS Z

BUS R
BUS T

Figure 23
Either of these conditions increases the zero sequence current
that flows in line 1 over the value that would be expected
without mutual induction. This increase in the zero sequence
current may cause a ground distance relay to overreach because of
the reduction in the apparent impedance. This is of particular
concern in the application of first zone ground distance relays
since they should not trip under these conditions. It may be
necessary to modify the relay settings to prevent this
undesirable operation. The percentage overreach for the
conditions of Case 1 and 2 are calculated in Appendix I.
When the currents in the protected line and the parallel line(s)
are flowing in the same direction the apparent impedance seen by
a ground distance function will be larger than it actually is as
cited in Case 3 below.
Case 3: All breakers are closed, a single line to ground fault
on bus Y.
The important consideration here is to be certain that an
overreaching ground distance relay used in a stepped distance or
pilot relaying scheme will operate for a fault on the remote bus,
with adequate margin. The results of Case 3 in Appendix I shows
that the apparent impedance seen by a ground distance function at

-16-
breaker a for an end-of-line fault is equal to 74 percent of the
actual impedance to the fault (a 26 percent underreach).
Therefore, to see this end of line fault, a ground distance
function at breaker A would have to have a setting greater than
135 percent of the actual line impedance. Sufficient margin
would have to be added to this setting to insure operation for
all line faults.
It is theoretically possible to compensate a mho ground distance
relay such that the effect of the zero sequence mutual on its
reach will be minimal. This is typically accomplished by
injecting a portion of the zero sequence current from the
parallel line into the relay to offset the increase or decrease
in the zero sequence voltage induced in the line. However,
mutual compensation may be undesirable or impossible for one of
the following reasons:

a) Where lines run parallel for a portion of their total

length, but do not terminate at the same substation at one
or both ends, it is not possible to obtain the zero sequence
current from the parallel line.

b) In what may be the most severe condition of zero sequence

mutual coupling, the case where the parallel line is open
and grounded at both ends, it is not possible to obtain the
zero sequence current from the parallel line for use in the
compensation circuit since the grounding point is usually
beyond the current transformer location.

c) The use of mutual inductance compensation may cause the

distance unit to lose directionality. This can occur on a
close in fault on the parallel line, where the zero sequence
compensation current may overpower the actual line currents
and allow the ground distance relay to operate for a reverse
fault.
EFFECT OF FAULT RESISTANCE

The effect of resistance in the fault is dependent on the type of

fault.
In faults not involving ground, the resistance is made up of the
arc products alone and it has been predicated that a constant
voltage (equal to about 5 percent of normal system voltage) is
maintained across the arc [12]. The resistance is therefore
non-linear and can be estimated to be equal to about 5 percent of
the total impedance from the fault back to the source behind the
fault (5 percent of ZSL as shown in Figure 24A). Phase distance
functions will be most affected by arc resistance for conditions
of high source to line impedance ratios (ZS/ZL) wherein the
resistance can appear to be large relative to the line impedance.

-17-
In faults involving ground, there is a linear component of
resistance in addition to the non-linear arc component. This
linear component can be introduced by tower footing resistance,
ground wire resistance, trees growing into the line, etc. For
ground faults, therefore, the voltage drop across the fault
resistance will be directly proportional to the total current in
the fault. In addition, infeed, load flow and non-homogeneity
(system impedance angles are different) will modify the effect of
the resistance on the relay performance.
Figure 24 will be used to demonstrate the effects of infeed. load
and non-homogeneity. An IR-IX diagram is used, which is the same
as an R-X diagram if the current is neglected.
For the system shown in Figure 24A, the voltage seen by the relay
at the left will be:

Let IR be the reference so that 8R = 0, then

v-R= IR·(ZL&& + ,IF/BF*RF) (5)
IR
ZR = VR = ZL& + SF/BF=RF (6)
IR IR

I 111

D. BREAKER A CLOSED, LEFT-TO-RIGHT LOAD. E. BREAUER A CLOSED, RIGHT-TO-LEFT LOAD.

ORDLcOR OR DL>OR

Figure 24 - Effect of load flow and system non-homogeneity on

fault resistance
-18-
If breaker A is open, the system is radial, and the relay
current, IR, is equal to and in phase with the current in the
fault resistance, IF. The fault resistance appears to be purely
resistive and it adds directly to the line impedance as shown in
Figure 1B.
If breaker A is closed, the current in the fault resistance, IF,
is greater than the relay current, IR, which causes the
resistance to appear larger than it actually is (by the ratio of
IF/IR). With no load flow and a homogeneous system, IR and IF
will be in phase and the resistance will be magnified but not
shifted in phase as shown in Figure 24C. Load flow or system
non-homogeneity causes IF and IR to be at different angles which
introduces an apparent reactive component to the magnified
resistance as shown in Figures 24D and 24E.
Exactly how a ground distance function will respond to resistive
faults is dependent on the type of function (reactance versus
mho) and on the method used to polarize the function.
Different methods of polarizing mho ground distance functions
were described earlier in this paper. As previously shown, a
healthy-phase polarized function produces an expanded
characteristic that is dependent on the source behind the
function. Load flow was not considered for those plots. When
load flow is accounted for, the characteristics are shifted as
shown in Figure 25.

ZL = LINE IMPEDANCE
z s = EQUIVALENT SOURCE IMPEDANCE

- - - a FAULT RESISTANCE FOR

INTERNAL FAULT

A. LEFT-TO-RIGHT LOAD FLOW B. RIGHT-TO-LEFT LOAD FLOW

Figure 25 - Effect of Load Flow on Mho Distance Function

-19-
Note that the characteristics shift in the same direction as the
shift in the fault resistance. For the external fault, there is
no tendency by the mho function to either underreach or overreach
regardless of the direction of the load flow. For internal
faults the shift is in the proper direction to provide optimum
protection for resistive faults.
Reactance functions can be affected similarly. The performance
of a phase-current polarized and a negative sequence current
polarized reactance function is shown in Figure 26.
For the system shown in Figure 24a, the following parameters were
used:
EL = 67 V /o' ER = 67 V /-30'
ZR = 6 n /88’ RF = .5 n 10’

Left Source Transmission Line Right Source

Z1S = 1.5 n /88' Z1L = 6 /88’
n Z1S = 1.5 n /88-
Z0S = 1.5 n jar Z0L = 18 n /88* zos = 1.5 n /aa*

Because the system impedances are all at the same angle (88
degrees), this is a homogeneous system. Note that the reach of
the functions is set equal to the line impedance and that the
fault is at the end of the line (a balance point fault). The
VPOL = P O L A R I Z I N G aJAN-rITY
IR = RELAY CURRENT (Il+I2+K0*I0)
ZR * R E L A Y REACH (POSITIVE SEQUENCE)
VR e: RELAY VOLTAGE (FAULTED PHASE)
VOP n OPERATING QUANTITY (IR*fR - VR)

\ IX

REACTANCE
\I VPOL = 12*X

JIFMF

A. PHASE-CURRENT POLARIZED B. NEGATIVE SEQUENCE CURRENT POLARIZED

Figure 26 - Effect of Left-to-Right Load Flow on Reactance

Functions

-20-
relay current is the compensated current (I1+I2+K0·I0). The
reactance functions shown will operate when the operating
quantity (VOP) is within 90 degrees of the polarizing quantity
(VPOL) .

A phase-current polarized function can overreach as demonstrated

in Figure 26A. Overreach comes about because the fault current
is shifted with respect to the phase current which is being used
to polarize the function.
The response of a negative sequence current polarized reactance
function is shown in Figures 26B. Note that the characteristic
shifts in the same direction as the fault resistance and that
there is no tendency to overreach such as displayed by the
phase-current polarized function. The shift in the
characteristic is always in the direction to provide optimum
coverage for internal faults with resistance. The overreach is
eliminated because the phase relationship between the negative
sequence current polarizing quantity and the total fault current
is not affected by load flow. If the system is not homogeneous,
the negative sequence fault current will be shifted relative to
the total fault current which will in turn cause a tendency for
the function to either overreach or underreach depending on the
impedance relationships. The shift in the fault resistance is
not as severe as caused by load and the functions can be designed
to eliminate overreaching.

RF

I
&- e0 RF FOR INTERNAL FAULT

A. PHASE-CURRENT POLARIZED B. NEGATIVE SEOUENCE CURRENT POLARIZED

Figure 27 - Effect of Right-to-Left Load Flow on Reactance

Functions
Figure 27 shows the effect of right-to-left load flow. Note that
the phase-current polarized function would underreach for these
conditions.
It has been shown that ground distance functions do provide
coverage for fault resistance with the amount of coverage being
-21-
dependent on the type and design of the function. However, they
can not be relied on to provide coverage for extremely high
resistance ground faults such as encountered for fires under the
line, trees growing into the line, etc. If these faults are of
concern, ground directional overcurrent functions should be used
to provide the required coverage.

TYPICAL PROBLEM AREAS FOR GROUND DISTANCE RELAYS

The discussion thus far has involved the performance ground
distance functions for single line to ground faults. Although
the ground distance functions are designed primarily for
detecting single line to ground faults, they may respond to other
fault types as well. The following are examples of fault and
system conditions that may cause ground distance relays to fail
to perform as expected. The mho units in the examples are
"self-polarized", but the problems may exist even if cross
polarization is used.
POSSIBLE LOSS OF DIRECTIONALITY ON EXTERNAL LLG
As discussed previously, the purpose of the zero sequence current
compensation factor in the operating quantity is to allow the
ground distance relay to measure the proper impedance to the
fault based,on the positive sequence impedance of the line.
However, this same zero sequence current compensation factor may
cause a misoperation of, the forward looking element on the
unfaulted phase during an external double line to ground fault.
Consider the system of Figure 28 with a BCG fault on the bus at
terminal B. Typical settings for the Zone 1 and Zone 2 ground
distance relays at B are shown in Table IV. The current and
voltage signals at Breaker 3 for the external BCG fault are shown
in Table V.

BUS A BUS B
PROTECTEO BCG
LINE

I -1
Z1 = 1.0 A T SO DEG Z1 =6.0 AT 61 D E G 21 q 6.0 AT 67 DEC

LO 4 . 0 A T 60 DEG 20 =18.0 AT 76 DEG ZO = 16.0 AT 76 DEG

El =67 V AT 0 DEG E2 = 67 V AT 0 DEG

EXTERNAL BCG FAULT

Figure 28

-22-
 TABLE IV
Zone 1 Zone 2
ZRl = 5.4 n ZRl = 9.0 n
ZRO = ZRl ZRO = ZRl

TABLE V
Sequence Phase
I1 = 5.5 A /94' IA = 0.08 A /-81'
12 = 4.1 A /-89' IB = 8.9 A l-12'
IO = I.5 A l-78' xc! = 8.2 A /-162'
V1 = 28.6 V f-1.6' VA = 85.7 V /-1.6'
V2 = 28.6 V f-1.6' VB = 0.0 v /o"
V0 = 28.6 V J-1.6' vc = 0.0 v /o'

Consider a simple Zone 2 phase A ground distance relay at Breaker

3 looking towards Bus A having an operating quantity of:

(IA - I0)·ZRl + K0·ZR0 - VAG (7)

The voltage phasors comprising the operating signal are shown in
Figure 29; The (IA -I0)·ZRl term is neglected because of its low
magnitude. Note that if the magnitude of the zero sequence
current term is larger than that of the restraint voltage VAG,
this unit will operate. For the conditions of this example, the
positive sequence impedance reach of the function would have to
be increased to approximately 29 ohms, or almost 500% of the
positive sequence impedance of the protected line.
VPOL

-VAG

Figure 29
-23-
The reach required to cause a misoperation for this example is
larger than would normally be set for the over reaching unit in a
channel scheme. However, consider the three terminal line
application shown in Figure 30 The currents and voltages at
Breaker 3 for the external BCG fault on bus B are shown in TABLE
VI.

BUS A PROTECTED BUS B

BCG
Z1 = 6 . 0 AT 87 DEG

to = la.0 AT 76 DEG #(-g-J

4 E2

Z 0 = 1.0 AT 80 DEG 20 =18.0 AT 76 DEG

El = 67 V At 0 DEG ?-
0 AT 88 DEG
Z1 = 1.0 6 E2rQVAtOWt

ZO-LOATISDLC

E2 =67 V AT 0 DEG

EXTERNAL BCG FAULT

3 TERMINAL LINE

Figure 30
TABLE VI
Sequence Phase
I1 = 26.6 A /93' IA = 0.3 A L-96’
12 = 19.1 A l-89' IB = 42.5 A /-13'
IO = 7.8 A j-81' IC = 40.0 A j-161'
Vl = 28.0 v /-1.3,' VA = 84.0 V /-1.3‘
v2 = 28.0 V l-1.3' VB = 0.0 v (0'
vo = 20.0 v /-1.3" vc = 0.0 v fl'

The voltage phasors comprising the operating signal are shown in

Figure 31 for a positive sequence reach setting of 6 ohms. Note
that magnitude of the zero sequence current term is larger than
that of the restraint voltage VAG, and therefore this unit will
operate. For the conditions of this example, the positive
sequence impedance reach of the function is only set to 100% of
the positive sequence impedance of the protected line which is
unrealistically short for an over reaching function. The choice
of polarizing signal for the ground distance relay will have
little effect on the performance for this situation; other means
must be used to prevent misoperation.

-24-
 VPOL

Figure 31

POSSIBLE ZONE 1 OVER REACH ON EXTERNAL LLG FAULT

Ground mho distance functions are subject to over reaching on
remote double line to ground faults for certain system conditions
and relay settings. Consider the system shown in Figure 32. A
phase B to C to ground fault, with a fault resistance of 0.5
ohms, is applied on bus B. The resulting fault currents and
voltages are shown in Table VII.
BUS A BUS B
RELAY
LOCATIOM PROTECTED n

Z1 = 1.5 AT 88 BEG 21 = 6.0 AT 87 BEG Ii2 - 1.5 AT Ia DLC

20 I: 1.5 AT 88 BEG to - 15.S At 75 OEG 20 t 1.5 AT 58 DEC

5% - 67 V AT 0 BEG Et=ITVATYTIODEG

Figure 32
TABLE VII
Sequence Phase

I1 = 7.9 A /-66° IA = 7.7 A j-37'

12 = 3.4 A /31' IB = 11.3 A /163'
IO = 0.8 A /87' IC = 6.1 A /39"

Vl = 56.2 V /-4.5' VA = 60.4 V l-8.6'

v2 = 5.1 V L-61' VB = 50.5 V /-124°
vo = 1.3 v l-4.5' vc = 58.0 V /119'

-25-
The Zone 1 mho ground distance functions are set as shown in
Table VIII. The Zone 1 is set to "see" ninety percent of the
line.

TABLE VIII
Zone 1
ZRl = 5.4 ohms
ZR0 = ZRl
z 1 ;;:
K0 =3
VOP = (11 + I2)·ZRl + KO*IO·ZRO - VAG
VPOL = VAG

For the above conditions, the operate and polarizing signals for
the Phase B ground distance function with self polarization are:
Operate : 11.2 V /-108'
Polarizing : 50.5 V /-124'
Coincidence: 164'
The operate and polarizing signals for the Phase C ground
distance function with self polarization are:
Operate : 20.1 V i-87'
Polarizing : 58.0 V 1119'
Coincidence: 26 °

The coincidence of these signals is sufficient to cause the phase

B function to operate, while the phase C function will not
operate. Thus the relay will over reach and trip Zone 1 for a
fault beyond the end of the line. In order to prevent the Phase
B function from operating, the reach of the Zone 1 must be
reduced to approximately 4.25 ohms, which is 71 percent of the
line. This condition will also cause phase selection problems in
single phase tripping schemes which use over reaching ground
distance functions as phase selectors.
SINGLE PHASE TRIPPING CONSIDERATIONS
INTEGRATED VS. SEPARATED SCHEMES
Single phase tripping schemes can be constructed in many
different ways utilizing different types of devices. However,

-26-
the basic approaches may be divided into two categories:
"Integrated Schemes" and "Separate Schemes".
Any single phase tripping scheme must be able to make two basic
decisions for each fault:
a. Whether or not the fault is within the trip zone
b. Which phase or phases are faulted.
An Integrated Scheme utilizes the same measuring functions to
perform both of the above, while a Separate Scheme would use one
set of functions to establish whether or not the fault is within
the trip zone and a separate set of functions to establish the
faulted phase(s).
If one opts for an integrated scheme, then the selection of
measuring units is restricted to those that can determine which
phase or phases are faulted, as well as whether the fault is
internal or external. For example, an Integrated Scheme could
not satisfactorily employ zero sequence directional overcurrent
relays for ground fault protection because by themselves these
relays can not determine which phases are faulted. Ground
distance relays are often used as the measuring functions in
Integrated Schemes, as they are essentially phase selectors as
well as measuring functions.
When one phase of a transmission line is open, an asymmetrical
condition exists on the system that results in negative and zero
sequence current flow throughout the power system which in turn
produce negative and zero sequence voltages throughout the
system. In addition, the relay voltage will be affected by the
location of the potential device: line side of the breaker or bus
side of the breaker. In any case, relay measuring functions that
are affected by zero or negative sequence quantities must be
designed to operate correctly during the open phase period.
POLARIZING VOLTAGE
When the ground distance functions are to be employed in an
Integrated Single Phase Tripping Scheme, the polarizing voltage
should be chosen such that the performance of the ground distance
functions on the sound phases is not impaired during the open
phase period. Figure 33 shows the effect of an open B phase on a
phase A ground distance function for the four polarizing voltages
discussed previously: this assumes no induced voltage on the open
phase.

-27-
 VA

/I
vc
VA

VI-0
L BO

FAULTBD PHASE QUADRATURE

)1t
VA
VM - VBWA + Anta
VA - 0.5 VC

vc VB-0

POSITIVE SEQUENCE
MEDIAN

Figure 33

The per-unit magnitude and phase angle of the polarizing voltages

for the conditions of Figure 33 are summarized below:

Faulted Phase: 1.000 /o'

Quadrature: 0.577 J+30'
Median: 0.880 /-19.1"
Positive Sequence: 0.667 /o'

The phase shift introduced in the quadrature and median polarized

functions would cause the ground distance function to respond
differently when a phase is open: for some system conditions,
this may result in either failure to see a fault, or over-reach
of a zone 1 element for an external fault. The faulted phase and
positive sequence polarized functions, on the other hand, will
respond properly whether or not a phase is open.

PHASE SELECTION
To simplify the logic required when distance functions are used
in an Integrated Single Phase Tripping Scheme, several criteria
must be met:
a. For single line to ground faults only the ground
distance function on the faulted phase should operate;
-28-
 it is also desirable that no phase distance function
operates for a single line to ground fault.
b. For phase to phase faults, no ground distance function
should operate.
C. For double line to ground faults, either two ground
distance functions should operate; or one phase
distance function should operate, preferably with no
ground distance functions.
In general, the source to line impedance ratio determines the
ease of obtaining correct phase selection in single phase
tripping applications. With large source to line impedance
ratios, phase selection is simply accomplished since it can be
shown that only the proper ground distance function responds to
single line to ground faults, only the phase distance functions
respond to phase faults, and the phase distance functions do not
respond to ground faults.
As the source to line impedance ratio decreases, the operation of
the distance functions is not as well behaved. For example, it
is well known that the phase distance functions will operate for
close-in single line to ground faults, thus requiring some form
of ground preference to prevent the phase distance function from
initiating a three phase trip. The problems of correct phase
selection are further aggravated by heavy load flow and, in some
cases, by a low ratio of zero sequence to positive sequence
source impedance.

BUS A BUS B
RELAY
L O C A T I O N F1 PROTECTED F2
LINE

,f
2 &J 4 E2

Zl= 1.5 AT 88 DEG Zl= 6.0 AT 87 DEG E2 = 0.3 AT 68 DEG

ZO= 1.5 AT 88 DEG ZD = 18.5 AT 76 DEG ZO q 0.12 AT 88 DEG

El = 67 V AT 0 DEG E2 = 67 V AT-60 DEG

Figure 34
The following examples are based on the system of Figure 34. An
over reaching ground distance mho function with a reach setting
of 12 ohms at 85 degrees is assumed in each of these cases.
1. Close in single line to ground fault.
This case assumes a phase A to ground fault directly in
front of the relay location, with a fault resistance of 3
-29-
 ohms. The operate and polarizing signals for the Phase C
ground distance function with memory action are:
Operate: 52.7 V f44.6'
Polarizing: 61.6 V J109.6'
Coincidence: 115°
The coincidence of these signals is sufficient to cause the
phase C function to operate. As the load flow increases,
the phase C function is more likely to operate.
2. Remote single line to ground fault
This case assumes a bolted phase A to ground fault at the
remote bus (bus B). The operate and polarizing signals for
a Phase C ground distance function with memory action:
Operate: 99.6 V J196.4'
Polarizing: 61.6 V /109.6'
Coincidence: 93.2"
Again, the coincidence of these signals is sufficient to
cause the phase C function to operate. As the load flow
increases, the phase C function is more likely to operate;
as the fault resistance increases, the phase C function is
less likely to operate.
3. Remote phase to phase fault.
This case assumes a bolted phase B to phase C fault at the
remote bus (bus B). The operate and polarizing signals for
the Phase B ground distance function with memory action are:
Operate: 95.1 V /268.4'
Polarizing: 61.6 V /229.6'
Coincidence: 141.2"
The operate and polarizing signals for the Phase C ground
distance function with memory action are:
Operate: 7.5 v J340.4'
Polarizing: 61.6 V /109.6'
Coincidence: 50.8"
The coincidence of these signals is sufficient to cause the
phase B function to operate, but the phase C will not
-30-
 operate. This may cause the relay system to incorrectly
trip only phase B for this fault.
In general, the severity of the problem of incorrect phase
selection in the above cases will be a function of the type and
shape of the relay characteristic and the settings employed. As
noted previously, an increasing source to line impedance ratio
will reduce the probability of having phase selection problems.
Of course, there are other cases that will also result in phase
selection problems.
SUMMARY

This paper has presented an overview of ground distance relays in

an effort to acquaint the reader with some of the design and
application considerations of these relays. The relays discussed
were simple two input designs; more complex designs are
available. Some of these complex designs will eliminate
application problems discussed in this paper. The examples used
in the paper are based on the fundamental frequency components of
the fault currents and voltages: that is, the effects of
transients, such as caused by CCVT's, series compensation, shunt
capacitance, saturated CT's etc., have been ignored.

-31-
 REFERENCES

[1] Wilkinson S.B. and Mathews C.A. Dynamic

Characteristics of Mho Distance Relays, GER-3742, a GE
publication.

[2] Alexander G.E.,Andrichak J.G.,Tyska W.Z. and Wilkinson

S.B., Effects of Load Flow on Relay Performance,
GER-3743, a GE publication

[3] Andrichak J.G., and Wilkinson S.B., Consideration of

Speed, Dependability, and Security in Pilot Relaying
Schemes, GER-3055 a GE publication

[4] Andrichak J.G., Polarizing Sources for Directional

Ground Relays, GER-3182, a GE publication

[5] Tyska W.Z., Polarization of Ground Distance Relays

[6] Keeney M.F., Mathews C.A., and Wilkinson S.B., Design

Considerations in the Development of a New Ground
Distance Relay, GER-3089, a GE publication

[7] SLY/SLYG 60/80 Phase and Ground Distance Relays,

GET-6651, a GE publication

[8] Review of Ground Relaying Concepts for Transmission

Line Protection; GET-6657, a GE publication

[9] Elmore W.A., Zero Sequence Mutual Effects on Ground

Distance Relays and Fault Locators, a paper presented
at Texas A&M, April, 1992

[10] Warrington A.R. van C., Graphical Method for Estimating

the Performance of Distance Relays During Faults and
Power Swings, AIEE Transactions, 68 (1949), PP608-620

[11] High Voltage Transmission Line Protection with Single

Pole Tripping and Reclosing, GET-6555, a GE publication

[12] Distance Relay Application Guide, GER-3199, a GE

publication

-32-
 APPENDIX I
CHANGE IN REACH OF GROUND DISTANCE
RELAYS CAUSED BY MUTUAL INDUCTION.

Depending on system conditions, the presence of mutual induction

between two parallel lines can cause a ground distance relay to
either overreach or underreach. Three cases representing
different conditions are described below.

BUS Y BUS Z

TYPICAL EXAMPLE OF MUTUAL COUPLING

BETWEEN PARALLEL LINES

Figure A
Case 1: Zero sequence current In a parallel line flowing in a
direction opposite to that in the protected line.

The condition to be considered for Case 1 is represented in

Figure A. The apparent impedance seen by the ground distance
relay is defined by the phase to neutral voltage at the relay
location divided by a current. Consequently, any change in the
voltage resulting from the mutual inductance, assuming no
corresponding change in current, will cause the relay reach to
change. For the situation in Figure A, the phase A to ground
voltage for a phase A to ground fault at location X can be
expressed as follows:
VAG = Il·Zl + I2·Z2 + I0·Z0 - I0'·Z0M (I-1)

Adding and subtracting I0·Zl, and substituting Zl for Z2, yields

-33-
the following:
VAG = Il·Zl + I2·Z2 + I0·Zl - I0·Zl + I0·Z0 - I0'·Z0M (I-2)
VAG = (11 + 12 + I0)·Zl +I0·(Z0 - Zl) -I0·Z0M (I-3)
VAG = IA·Z1 + I0·(Z0 - Zl) - I0'·Z0M (I-4)
The percent change in reach, in this case overreach, when mutual
compensation is not employed can then be represented by the
following:
percent IA l Z1 IO fZ0 - Z1) Xl00 (l-5)
overreach = IA·Zl + IOT(Z0' - Z1) - I0'·Z0M

Assume the following system conditions as a most severe cause and

refer to Figure A for system configuration:
1. Line constants of the double-circuit 345 KV tower shown
in Figure 2 are:
Zl = Zl' = 0.061 +j 0.588 ohms/mile (0.038 + j 0.363 Ohm/~)

z o = Z0’ = 0.549 tj 1.964 ohms/mile (0.340 +j 1.211 ohms/km)

ZOM -0.489 tj 1.3 ohms/mile (0.302 tj 0.802 ohms/km)

2. Line R-T parallels line X-Y for 100 percent of line

length.
3. The source behind breaker H contributes 50% of IO in line
X-Y, and the source behind breaker L contributes the other 50%
over line T-R.
4. The ratio of zero sequence current to total line current
is 1:3.
For these conditions:

IO = IA/3 (I-6)
10' =IA/6 (I-7)

Substituting these Current values and the impedance values into

equation (l-5) yields an overreach of approximately 25 percent.
That is a relay set to reach 68 percent of the impedance of the
protected line will now reach 85 percent of the line.

-34-
Case 2: Grounded parallel line.
The conditions to be considered in Case 2 are as follows:
1 Line R-T is out of service with three-phase grounding
chains at both ends.
2. The ratio of zero sequence current in the line to total
line current is 1:3.
3. Impedances are the same as in Case 1.
The phase to neutral voltage on the'faulted phase at breaker A
is:
VAG = IA·ZL + I0·(Z0 - Zl) - IO*ZM*/ZO' (I-8)
The percent overreach is determined by the following equation:
(I-9)
percent IA·ZL + I0·(Z0 - Zl)
-1 Xl00
overreach = IA·ZL + I0·(Z0 - Zl) - IO*ZM*/ZO'

Substituting the current and impedance values into equation (l-9)

yields an overreach of approximately 38%.
Thus, a zone-l ground distance relay, normally set for 80 percent
of the protected line, would have to be set for 58 percent of the
line in order to maintain the same margin.
Case 3: Zero sequence current in a parallel line flowing in the
same direction as that in the protected line.
The condition for Case 3 would be represented by the system in
Figure A with all breakers closed, and with an external fault on
line 3. For this condition IO = IO', and assume that IO = IL/4.
The [Link] neutral voltage on the faulted phase at breaker A
is:
VAG = IA·Zl + I0·(Z0 - Z1) + I0'·Z0M (I-10)

The percentage underreach can then be expressed as follows:

(I-11)
percent Il·ZL + I0·(Z0 - Z1) Xl00
underreach = Il·ZL + I0·(Z0 - Zl) + I0'·ZM -'

-35-
Substituting,the current and impedance values into equation (I-
11) indicates a percentage of underreach of 26 percent. Thus, a
zone-2 relay set to reach 135 percent of the protected line would
actually reach just 100 percent of the line for the conditions
assumed.

-36-