Understanding the Dynamic Mho

Distance Characteristic

Donald D. Fentie
Schweitzer Engineering Laboratories, Inc.

© 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained
for all other uses, in any current or future media, including reprinting/republishing this material
for advertising or promotional purposes, creating new collective works, for resale or
redistribution to servers or lists, or reuse of any copyrighted component of this work in other
works.

This paper was presented at the 69th Annual Conference for Protective Relay Engineers and can
be accessed at: https://doi.org/10.1109/CPRE.2016.7914922.

For the complete history of this paper, refer to the next page.
 Revised edition released September 2023

Previously revised edition released April 2017

Previously presented at the

52nd Annual Minnesota Power Systems Conference, November 2016,
and 69th Annual Conference for Protective Relay Engineers, April 2016

Originally presented at the

42nd Annual Western Protective Relay Conference, October 2015
 1

Understanding the Dynamic Mho

Distance Characteristic
Donald D. Fentie, Schweitzer Engineering Laboratories, Inc.

Abstract—In order to retain dependability and security in cases

of close-in faults when the loop voltage is zero, mho distance Relay
elements use cross-phase and/or memory polarization. ES ER
Polarization techniques in distance elements cause the mho
(a) ZS1 mZL1 (1 – m)ZL1 ZR1
distance characteristic to dynamically expand and then contract
as the polarization memory decays during a fault. The dynamic
behavior of this mho element continues to be a confusing topic to RF
many engineers. Many papers and text books written on the
subject are mired in long derivations and produce equations that
can seem abstract.
This paper bridges the theory with familiar phasor diagrams
and visual aids to promote an intuitive understanding of the X
dynamic mho expansion during balanced and unbalanced faults in
the forward and reverse direction. We examine the dynamic mho 3
expansion for common fault types, load flow conditions, and under
Zr
typical relay test scenarios. We emphasize visualizing the
expansion for positive-sequence memory polarization while 2
demonstrating its practical benefits in terms of security, (b)
directionality, and increased resistive coverage. Z

1
I. INTRODUCTION m ∙ Zr Operating Restraining
Mho distance elements continue to be popular for
transmission protection worldwide. In the past, as relay R

technology improved from coils and discs to microprocessors, –2 –1 1 2

improvements were made to algorithms, polarization, and

response times. The fundamental principles of mho distance Fig. 1 Line fault at 50 percent of reach (m = 0.5 p.u.): (a) System one-line
elements still apply, and there is continued interest in diagram, (b) Mho characteristic.
understanding this seemingly simple protective element.
If we assume no arc impedance, the calculated impedance
This paper discusses the mho element and dynamic
for a fault at the relay is zero. If the measured impedance is
expansion characteristics for positive-sequence polarization
plotted for a three-phase fault at each point along the
during testing, load flow, and fault conditions. It also discusses
transmission line, it will produce a bolted fault locus with the
element security during single-pole open and series-capacitor
same angle as the transmission-line impedance. At a certain
applications. This paper departs from many other papers that
distance from the relay, Z equals the mho reach Zr. All points
present mathematical equations governing mho expansion by
on the bolted fault locus between the relay reach point and the
instead emphasizing visualization to promote a more intuitive
origin are enclosed within the area of operation.
understanding [1] [7] [9] [10].
A. The Self-Polarized Mho Circle
II. BACKGROUND Microprocessor relays require an efficient method for
A self-polarized mho element for the system shown in determining if a measured impedance is inside the mho
Fig. 1(a) is represented as a circle on the impedance plane in operating characteristic. One technique is to examine known
Fig. 1(b). The diameter of this circle extends from the origin of phase angle relationships, as shown in Fig. 2. Note that the
the plane to the relay reach setting Zr on the bolted fault locus. measured impedance Z in this example does not lie on the
The relay calculates the impedance Z by using the measured bolted fault locus because of arc resistance.
voltage and current phasors at each processing interval. The
mho element restrains until Z plots inside the circle.
 2

All other angle differences signify that the impedance is outside

the mho circle. These three conditions are shown in (2), (3), and (4).
X
Operate (inside mho):
3
 dZ 
−90° < arg  < 90° (2)
Zr  Z 
dZ
2 Restrain (outside mho):
(a)
 dZ 
δ < 90° 90° < arg  < −90° (3)
Z
 Z 
1
Indeterminate (on the mho circle):
 dZ 
R arg   =±90° (4)
–2 –1 1 2
 Z 

X 3 dZ1
δ1
3 δ = 90°
dZ2 Zr
Zr δ2 2
dZ
Z1 dZ3
2 δ3
(b) Z2
Z 1 Z3

R
–2 –1 1 2
R
–2 –1 1 2

Fig. 3 Right-angle property of a circle.

The impedance from the relay to the reach point determines

δ > 90° the diameter of the mho circle. Vectors dZ and Z will always be
X
at right angles to each other when the impedance point lies on
dZ
3 the mho circle. The angle comparison is typically shown in
terms of measured voltage by multiplying the numerator and
Zr
denominator in (4) by the faulted loop current I, as shown in (5).
2
(c)  dZ   Zr − Z   I ⋅ Zr − V 
arg   arg
= =  arg   (5)
Z  Z   Z   V
1 Although (5) is mathematically correct, it is not a
computationally efficient representation. We can simplify the
equation by using comparators, as discussed in the next subsection.
R
–2 –1 1 2 B. Comparators
A phase comparator determines if the angle difference
Fig. 2 Impedance vectors for faults (a) Inside zone, (b) At reach point, between two phasors is within a specified range [1]. In addition
(c) Outside zone. to (5), the angle difference between dZ and Z can be represented
as shown in (6).
Vector dZ defines the impedance between the fault and the
relay reach point as shown in (1). (
arg dZ ∠θ dZ ⋅ Z ∠θ Z * = θ dZ − θ Z ) (6)
dZ
= Zr − Z (1) A measured impedance is inside the mho circle when the
When the measured impedance is inside the mho circle, the result of (6) is between −90 and +90 degrees. The cosine
angle δ, defined as θdZ − θZ, is between −90 and +90 degrees. function with no phase shift is positive between −90 and +90
The diameter of the circle is defined by Vector Zr, which splits degrees, and thus it makes an excellent comparator for
the mho into a left and right side. When the impedance lies determining whether the impedance is within the mho element.
directly on the edge of the left half or right half of the mho, If the cosine of the angle difference between dZ and Z is
δ equals −90 and +90 degrees, respectively, as shown in Fig. 3. positive, the impedance is inside the mho element.
 3

We can use Euler’s formula to rewrite the expression inside Re  Z ⋅ Z * 

the brackets of (6), separating it into real and imaginary Zr = (15)
components, as shown in (7). Re 1∠θ Zr ⋅ Z * 

dZ ⋅ Z *= (
dZ ⋅ Z ⋅ cos (θ dZ − θ Z ) + j ⋅ sin (θ dZ − θ Z ) ) (7) Equations (16) and (17) show the circle-mapping technique
in terms of voltage:
Taking the real part of (7) yields the cosine comparator (8),
the result of which is positive if the impedance plots inside the Re ( Z r ⋅ I − V ) ⋅ V *  =0 (16)
mho circle.
Re V ⋅ V * 
Re  dZ ⋅ Z *  = dZ ⋅ Z ⋅ cos (θ dZ − θ Z ) > 0 (8) Zr = (17)
Re 1∠θ Zr ⋅ I ⋅ V * 
Substituting (1) into (8) gives (9):
Renaming Zr as Zmapped in (15) yields (18):
Re ( Z r − Z ) ⋅ Z *  > 0 (9)
Re  Z ⋅ Z * 
Equation (9) is often expressed in terms of voltage, as shown Z mapped = (18)
Re 1∠θ Zr ⋅ Z * 
in (10), by multiplying each term by current, where V is the
measured voltage at the relay and I is the fault loop current, as Zmapped is the magnitude of the point of intersection between
described in Section III, Subsection A [1]. the dotted circle and the bolted fault locus, as shown in Fig. 5.
This impedance is analogous to the per unit quantity r that is
Re ( Z r ⋅ I − V ) ⋅ V *  > 0 (10)
mapped on the M-line [1]. In Fig. 5(a), Zmapped is positive and
The mho element in Fig. 4 is represented in terms of voltage less than the reach point Zr, so the element operates. In Fig. 5(b),
on the complex plane. Zmapped exceeds the reach point Zr, so the element restrains.

I∙X δ X
Impedance
3
50 mapped to
dV line angle
Zr
40
I ∙ Zr 2 dZ

30
(a)
V Zmapped 1
20

Z
10
R
–2 –1 1 2
I∙R
–20 –10 10 20 30 Zmapped Zr

0 1 2 3
Fig. 4 Mho element in terms of voltage.

We can improve (9) so that it projects or maps Z to the bolted Impedance

X mapped to
fault locus as a scalar quantity that we can then compare 4 line angle
directly with a scalar reach point on a number line [1]. By
setting (9) equal to zero, as shown in (11), we define a circle
that passes through the origin and Z. The diameter of the circle 3
dZ

lies along the line angle. Zr

Re ( Z r − Z ) ⋅ Z *  =
0 (11)
(b) Zmapped
We can solve (11) for the magnitude of Zr to find the mapped
impedance, as shown in (12), (13), and (14). Z

(
Re  Z r ∠θ Zr − Z ⋅ Z *  =
0 ) (12)

Re  Z r ∠θ Zr ⋅ Z *  − Re  Z ⋅ Z *  =
0 (13) R
–2 –1 1 2
Re  Z ⋅ Z *  Zr Zmapped
Zr = (14)
Re 1∠θ Zr ⋅ Z *  0 1 2 3 4

Zr can also be mapped as a negative number in (14) by Fig. 5 Mapping a measured impedance to the line angle and corresponding
removing the absolute value sign, as shown in (15). number line for (a) A fault within reach point, (b) An external fault.
 4

Equation (18) benefits us by reducing the number of Extensive research of polarization has resulted in the
comparator calculations to just one per relay-processing examination of several solutions [1] [2] [6] [7]. Cross-
interval, regardless of the number of mho zones. The output of polarization uses the unfaulted phases during a fault. For
the equation Zmapped is a scalar quantity that we can compare example, a BC mho element might use a shifted voltage vector
with the scalar mho reach points of all other zones. This method of VA for polarization instead of VBC. Also, an AG mho element
is more computationally efficient than performing (9) for each might use a shifted voltage vector of VBC for polarization
mho element. instead of VA. Positive-sequence voltage is another polarization
type that produces excellent results by using all three phases.
III. POLARIZATION The polarization techniques previously discussed are
unreliable for three-phase faults close to the relay because the
A. Polarization Choices
voltage on each phase is zero. One solution to this problem is
The term Zr ∙ I − V in (16) is defined as the operating quantity to use a memory filter in the polarization algorithm so that the
dV that can change by 180 degrees, depending on fault location, voltage decays slowly over many cycles [2]. This technique
as shown in Fig. 6. We then compare the operating quantity dV provides a way of dealing with a bolted three-phase fault at the
to the polarizing quantity, which is defined as the measured relay location for a period of time, depending on the memory
voltage V in (16). decay.
The polarizing quantities for self-polarizing and positive-
I∙X
sequence memory polarizing are shown in Table 1 for the six
50 fault loops.
V
45 TABLE 1
OPERATING AND POLARIZING QUANTITIES
40
Self- Positive-Sequence
35 dV Relay
Polarizing Memory Polarizing Loop Current
Element
30 Quantity Quantity
(a) I ∙ Zr
25 AG VA VA1mem IA + k0 ∙ IRES

20 BG VB VB1mem IB + k0 ∙ IRES

15 CG VC VC1mem IC + k0 ∙ IRES

10 AB VAB VAB1mem IA − IB

5 BC VBC VBC1mem IB − IC
I∙R CA VCA VCA1mem IC − IA
–15 –10 –5 5 10 15 20
We can replace the self-polarizing voltage V in (17) with the
positive-sequence memory voltage V1mem, as shown in (19). The
I∙X numerator of (19) still contains the term V that was derived
60 from the operate quantity dV.
dV
50 I ∙ Zr Re V ⋅ V1mem* 
Zr = (19)
40 Re 1∠θ Zr ⋅ I ⋅ V1mem* 
(b) V
30 During a fault, the self-polarizing voltage at the relay
20 decreases on the faulted phases while the memory voltage
slowly decays from prefault levels. This causes the memory-
10 polarized mho circle to become larger than the self-polarized
I∙R circle because the angle comparison is now between Vp (V1mem)
–30 –20 –10 10 20 30 40 and dV, as shown in Fig. 7(b). The impedance equivalent of Vp
is the polarizing impedance Zp, as shown in Fig. 7(a). Even
Fig. 6 Operating quantity dV direction is determined by fault location for an
(a) External fault, (b) Internal fault.
without memory, positive-sequence voltage polarization causes
the mho to expand. An exception is in the case of a close-in,
The angle of the polarizing term should remain relatively three-phase fault because the voltage is zero on all three phase.
constant during a fault. We call the mho element in Section II The memory voltage decays over time, causing the dynamic
the self-polarized element because the relay uses the voltage of circle to shrink. The steady-state resting point of the dynamic
the faulted phase(s) for polarization. This polarization circle depends upon several system parameters and the fault
jeopardizes the security and dependability of the element during type, as discussed in Section III, Subsection B. One benefit of
close-in faults because the voltage is too small to be accurately the dynamic expansion is the initial increase in resistive
measured. Modern relays use other forms of polarization that coverage provided by this expansion, without changing the
perform reliably for all fault types and locations. reach point.
 5

From Table 2, we can observe that the prefault current IPrefault

affects the positive-sequence voltage at the relay the moment
X
before the fault occurs. Section III, Subsection C explores the
3 effects of load flow.
Zr B. Visualizing Positive-Sequence Polarization
dZ
2 We can use a two-source system to analyze a typical relay
response on a transmission line, as shown in Fig. 1(a). The
(a) Z
system parameters for the following examples are kept simple
1 in order to aid visual understanding, as shown in Table 3.
TABLE 3
Zp
VISUALIZATION SYSTEM PARAMETERS
R
–2 –1 1 2 Parameter Value Unit Description
ZS1 1∠80° Ω (sec) Sending-end positive-sequence
source impedance
ZS0 3∠80° Ω (sec) Sending-end zero-sequence
source impedance
ZL1 3∠80° Ω (sec) Positive-sequence line impedance

I∙X ZL0 9∠80° Ω (sec) Zero-sequence line impedance

70
ZR1 1∠80° Ω (sec) Receiving-end positive-sequence
60 I ∙ Zr source impedance
dV
50 ZR0 3∠80° Ω (sec) Receiving-end zero-sequence
40 source impedance

(b) 30 V ES 67∠0° V (sec) Sending-end source voltage

20 ER 67∠0° V (sec) Receiving-end source voltage

10 Vp Z1P 3 Ω (sec) Relay reach setting
I∙R
–50 –40 –30 –20 –10 10 20 30 40 50 k0 0.67∠0° unitless Zero-sequence compensation
–10 factor
ZFG* 0.5 Ω (sec) Ground fault impedance
ZFP* 0.6 Ω (sec) Phase-to-phase fault impedance
m 0.43 p.u. of Distance on line to fault
Fig. 7 Dynamic mho characteristic on the (a) Impedance plane,
ZL
(b) Voltage plane.
* Where applicable.
Positive-sequence memory polarization may be preferred
over cross-polarization because it improves single-pole trip The maximum mho expansion occurs at fault inception
security and has excellent expansion properties [1]. The because V1mem equals the positive-sequence voltage just prior to
positive-sequence polarizing quantities and polarizing the disturbance. The memory voltage diminishes over time until
equations for each relay element in Table 1 are shown in it reaches the final value of positive-sequence voltage during
Table 2. the fault. The following subsections describe this expansion and
TABLE 2
decay for each type of fault.
POSITIVE-SEQUENCE MEMORY POLARIZING QUANTITIES*; a = 1∠120° 1) Three-Phase Fault
Polarizing Quantity Polarizing Equation Think of the origin of the impedance plane as the location of
the relay. During a fault on the line, the measured value of the
VA1mem ES − IPrefault ∙ ZS1
faulted voltage V diminishes, while V1mem initially retains the
VB1mem a2 ∙ VA1mem value it had before the disturbance. On the impedance plane,
VC1mem a ∙ VA1mem this means that the magnitude of the measured impedance Z is
VAB1mem −j ∙ a ∙ VA1mem ∙ 3
significantly less than the polarizing impedance Zp, and the
dynamic mho circle envelops the origin, as shown in Fig. 8.
VBC1mem −j ∙ VA1mem ∙ 3
VCA1mem −j ∙ a ∙ VA1mem ∙
2
3
* Based on the system shown in Fig. 1(a); ABC rotation assumed.
 6

X X

3 3

Zr Zr

2 dZ 2 dZ

Zp
1 Z 1 Z

Zp
R R
–2 –1 1 2 3 –1 1 2
ZS1

Fig. 9 Three-phase fault dynamic mho characteristic (steady state).

Vectors Z and Zp overlap.

Fig. 8 Dynamic mho characteristic (maximum expansion).

40
VC
We calculate the polarizing impedance Zp by dividing V1mem IC
by the loop current. If we ignore the effects of load flow, V1mem IB
VA
initially equals the source voltage ES; therefore, in this case,
–40 40
Zp equals the source impedance plus the apparent impedance
measured by the relay to the fault Z. This is shown graphically IA
VB
in Fig. 8, where the impedance from the relay location at the –40
origin to the tail of Vector Zp equals Zp − Z = ZS1.
The mho expansion for forward faults depends on the
magnitude of the source impedance, a concept known Fig. 10 Three-phase fault vectors for voltage and current.
commonly as “expanding back to the source.” The greater the
ratio of the source impedance to the line impedance (source-to- 60
impedance ratio, or SIR), the greater the dynamic mho
VC1mem 40
expansion. Note that the maximum expansion back to the source
shown in Fig. 8 uses the same system parameters and applies
for all types of faults examined in this section, unless otherwise VA1mem

noted. –60 –40 40 60

The memory voltage V1mem eventually decays to the
measured positive-sequence voltage V1, which is equal in VB1mem –40
magnitude to VA, VB, and VC in a balanced system; therefore, the
dynamic mho circle for a three-phase fault eventually shrinks –60

to the smaller self-polarized circle, as shown in Fig. 9. The

voltage and current vectors for this diagram are shown in
Fig. 11 Three-phase fault relay vectors; dashed lines represent expanded
Fig. 10. Fig. 11 shows the substantial voltage decay over time memory voltage vectors, and solid lines represent steady state.
for a three-phase fault.
 7

Based on [3], the expansion is derived mathematically as a The expansion properties of the dynamic mho circle during
vector that begins at the origin and ends at the dynamic mho phase-to-phase faults are simple, for two reasons. First, the
circle. This vector is represented as b for the steady-state mho derivation does not involve the zero-sequence network and so
characteristic, and bmem for the full expansion. For a three-phase does not contain zero-sequence impedances. Second, negative-
fault, we have (20). sequence impedances are often assumed to equal corresponding
b=0 positive-sequence impedances.
(20)
bmem = − Z S1

X 3

Zr
3

2 dZ
Zr

1 Z

1
Zp
R
–2 –1 1 2
R
–2 –1 1 2 0.5 ∙ ZS1
b(mem) –1

Fig. 13 Phase-to-phase fault dynamic mho characteristic (steady state).

Fig. 12 Vector bmem representation of expansion and steady state.

VC
40
In this example, bmem, shown in Fig. 12, has a magnitude of
ZS1. When the memory voltage reaches steady state, b is at a IC
distance of zero from the origin, which means that the dynamic VA

circle and the self-polarized (or “static” circle) overlap. –40 40 60

IB
2) Phase-to-Phase Fault
The dynamic mho for a phase-to-phase fault initially –40
VB
expands back to the source (as explained in Section III,
Subsection B) and eventually decays to 0.5 ∙ ZS1, as proved by
Calero [3]. We can express the dynamic expansion vector as Fig. 14 BC fault vectors for voltage and current.
shown in (21).
Z S1 60
b= −
2 (21) VC1mem
40
bmem = − Z S1
The dynamic mho does not shrink to the self-polarized mho VA1mem
circle, no matter how much time elapses, as shown in Fig. 13. –40 –20 20 40 60 80
This is logical because the polarizing memory voltage VBC1mem
at steady-state fault current exceeds the measured quantity
VB − VC (see Table 2). On the impedance plane, this means that VB1mem –40
the polarizing impedance Zp always exceeds the measured loop –60
impedance Z. The voltage and current vectors for the phase-to-
phase fault are shown in Fig. 14. The memory voltage decay for
a phase-to-phase fault is less than that for a three-phase fault, Fig. 15 BC fault relay vectors; dashed lines represent expanded memory
as shown in Fig. 15. voltage vectors, and solid lines represent steady state.
 8

3) Phase-to-Phase-to-Ground Fault 60
As with other faults we have examined up to this point, the
dynamic mho expansion during a phase-to-phase-to-ground VC1mem 40

fault depends on the source impedance ZS1. The dynamic mho

characteristic decays to a value given by (22), as proved by VA1mem
Calero [3]. –40 –20 20 40 60

 Z S0 + m ⋅ Z L0 
b = − Z S1  
 Z S1 + m ⋅ Z L1 + 2 ⋅ Z S0 + 2 ⋅ m ⋅ Z L0  (22) VB1mem –40

bmem = − Z S1 –60

The derivation of non-memory Vector b is significantly

more complicated than for three-phase or phase-to-phase faults. Fig. 18 BCG fault relay vectors; dashed lines represent expanded memory
The expansion depends on the zero-sequence network as well voltage vectors, and solid lines represent steady state.
as on the impedance of the line to the fault in per unit m. The 4) Phase-to-Ground Fault
maximum decay, shown in (22), will be less than 0.5 ∙ ZS1, as The expansion of the phase-to-ground dynamic mho is
shown in Fig. 16. The voltage and current vectors for this fault influenced by the current compensation factor k0 [6]. The
are shown in Fig. 17. The memory voltage decay, shown in expansion vectors are shown in (23) [3].
Fig. 18, is similar to a phase-to-phase fault.
 Z S0 
 1+ Z 
b = − Z S1  S1

X
 2 + Z L0 
 Z L1 
3 
(23)
Zr  Z S0 
2+ Z 
2 dZ bmem = − Z S 1  S1

 2 + L0Z 
 Z L1 
 
1 Z The maximum expansion described by bmem in (23) has a
magnitude of ZS1 in systems where the angles and ratios of
Zp
R
{ZS0, ZS1} and {ZL0, ZL1} are equal. The expansion shrinks to
–2 –1 1 2
0.80 ∙ ZS1, as shown in Fig. 19, using the system parameters in
Table 3. The dynamic characteristic for a phase-to-ground fault
0.43 ∙ ZS1
provides the benefit of increased resistive fault coverage, which
–1 is often crucial during ground faults. The voltage and current
vectors for this fault are shown in Fig. 20. The memory voltage
Fig. 16 Phase-to-phase-to-ground fault dynamic mho characteristic
decay for a phase-to-ground fault is shown in Fig. 21. The
(steady state). steady-state positive-sequence voltage remains strong because
of the two healthy phase voltages.
40
VC
X
IC
IB VA 3

–40 –20 20 40 60
Zr

VB 2 dZ
–40

Fig. 17 BCG fault vectors for voltage and current. 1 Z

Zp
R
–2 0.8 ∙ ZS1 1 2

–1

Fig. 19 Phase-to-ground fault dynamic mho characteristic (steady state).

 9

60 X
VC
40 4

VA
3
–60 –40 –20 20 40
Zr
IA (a)

–40 2

VB Zp ZL1 + ZR1
–60

Fig. 20 AG fault vectors for voltage and current. dZ

R
–2 –1 Z 1 2
60
VC1mem
40

X
VA1mem
3
–40 –20 20 40 60
Zr

–40 2
VB1mem (b)
–60 Zp 0.8 ∙ (ZL1 + ZR1)

Fig. 21 AG fault relay vectors; dashed lines represent expanded memory dZ

voltage vectors (not visible in this figure because the vectors are too short),
and solid lines represent steady state. R
–2 –1 Z 1 2
5) Reverse Fault
We can simulate a reverse fault by shifting the fault current Fig. 22 Close-in reverse phase-to-ground fault dynamic mho characteristic
phasors by 180 degrees; the voltage change is negligible for a with arc resistance added: (a) Maximum expansion, (b) Steady state. Note that
close-in fault on either side of the current transformers (CTs). the dynamic mho (dashed circle) determines operation. The self-polarized
For a close-in reverse fault in Fig. 1(a), the relay measures fault mho (solid circle) is shown as a reference.
current supplied by the remote source ER. The relay source Several impedance vectors change direction for a reverse
impedance is now in front of the relay, with respect to the fault fault, causing the dynamic mho to shrink. Vectors Z and Zp
location, and equal to ZL1 + ZR1. reverse directions for faults behind the relay; Zr is constant in
The source impedance begins in the first quadrant on the the relay defined by the reach setting, and it continues to point
impedance plane and ends at the origin for an inductive system, in the forward direction. The polarizing impedance Zp points
as shown in Fig. 22. This represents the impedance from the from the location of the source to the measured impedance Z.
remote source to the relay location. Adding a small amount of Because the dynamic mho is defined between the tail of Zp and
fault resistance separates the vectors and allows them to be seen the tip of dZ, the dynamic circle becomes smaller than the self-
more clearly. polarized mho under this condition.
 10

A smaller operating characteristic during reverse faults

alleviates some directional security concerns regarding the mho Im(V)
expansion. In Fig. 22(b), the distance between the origin and
40
the tail of Vector Zp initially equals | ZL1 + ZR1 |. At steady state,
ES
Zp shrinks to 80 percent of its original value, Fig. 22(b), using ES − Ipf ∙ ZS1
the parameters in Table 3. Over time, the decay in the
polarizing quantity is analogous to the forward AG event (b)
VA1mem
described under “Phase-to-Ground Fault” in Section III, 20
Subsection B. We can apply similar logic to other reverse fault
types.
C. Load Flow Re(V)
In a two-machine lossless system, such as in Fig. 1(a), with 20 40 60
no fault, we can represent the power sent over the transmission
line by the power-flow equation shown in (24).

P1,2 =
ES ⋅ ER sin δ S , R( ) (24)
Im(V)

X total 40

where
δ S , R arg ( ES ) − arg ( ER )
= (25)
(c)
Xtotal is the total reactance between sources ES and ER. Power 20
VA VA1mem
flows from the source with the leading voltage angle, and this
creates a positive-sequence voltage drop across the source
impedance; therefore, the positive-sequence memory voltage at
the relay is unequal to the source voltage at fault inception. The Re(V)

memory voltage phase angle may shift, changing the position 20 40 60

of the polarizing impedance Zp.
The voltage drop across the source impedance is relatively Fig. 23 Voltage vectors during a three-phase fault with prefault current
(positive power flow) on a homogeneous system with no fault resistance;
small during prefault conditions when compared to a bolted compare (a) ES with VA, (b) ES with VA1mem, (c) VA with VA1mem.
fault. We simulate this in Fig. 23, which shows a three-phase
fault placed on a homogeneous system with positive power If we define Ifault as the fault loop current, we calculate the
flow and no fault resistance. VA1mem represents the prefault two impedance vectors as shown in (26) and (27).
(memory) voltage, and VA is the measured fault voltage on V1mem
Zp = (26)
A-phase. I fault
VA
Im(V) Z= (27)
I fault
40
ES If we extend the logic of (26) and (27) to all forward faults,
Vectors Zp and Z are separated by angle θ, as shown in Fig. 24.
For example, during an AG fault, θ represents the phase shift
(a) between VA and VA1mem. Over time, θ decreases as the memory
20
dVA voltage decays.
VA

Re(V)
20 40 60
 11

IV. TESTING
X
A. Justification for Testing
3 Dynamic mho expansion varies based on manufacturer
Zr
dZ implementation of the elements. In one implementation, the
time constant of the memory polarization is four cycles [2].
2 We advise testing the dynamic mho when assessing the
θ performance of the element on lines with series capacitors. The
Z dynamic mho expansion is necessary to detect forward faults
1
that are directly in front of a line-end capacitor with bus-side
Zp potential transformers (PTs) (see Section V, Subsection B).
R
Often, this type of testing employs a real-time simulator, where
–2 2 the relay is included in the playback loop.
In practice, the static or self-polarized mho is tested
extensively at steady state; for most applications, the reach
–1
point is the most important aspect of the element. We
recommend testing mho elements by using values from a
Fig. 24 Positive load flow reducing resistive fault coverage by shifting the constant-source impedance model.
dynamic mho to the left.
B. Testing Considerations
Vector VA1mem, and therefore Zp, increasingly lags the
Section III shows that the dynamic circle does not shrink
measured fault voltage VA for forward power flow as the
back to the self-mho circle except in the case of a three-phase
prefault current increases. This causes Zp to shift the dynamic
fault. This may be confusing to a relay tester using a common
mho circle to the left for forward load flow; reverse load flow
test procedure because the mho element appears to operate as if
has the opposite effect. Four impedance graphs in Fig. 25
the dynamic circle did not exist.
illustrate this effect for different levels of load flow.
The dynamic mho element does not usually interfere with
testing because of the test voltages that are typically used.
X X
Table 4 lists several test voltages and currents used to test a
3 3
Zr Zr
phase-to-ground mho circle by simulating an AG fault.
2 dZ 2 dZ TABLE 4
(a) (b) EXAMPLE TEST QUANTITIES
Z Z
1 1
Zp Zp
R R Test
Test 1 Test 2 Test 3 Test 4
–2 1 2 –2 1 2 Quantity
VA V (sec) 27∠0° 27∠0° 27∠20° 27∠−20°
VB V (sec) 67∠−120° 67∠−120° 67∠−120° 67∠−120°
X X VC V (sec) 67∠120° 67∠120° 67∠120° 67∠120°
3 3
Zr Zr
V1 V (sec) 53.67∠0° 53.67∠0° 53.21∠3.3° 53.21∠−3.3°
2 dZ 2 dZ IA A (sec) 3∠−20° 1.6∠−60° 2.64∠−20° 3.19∠−20°
(c) (d)
Z Z
1 1 Note: IB = IC = 0 A.
Zp Zp
R R
The voltage angles in Tests 1 and 2 for VA, VB, and VC are,
–2 1 2 –2 1 2
respectively, 0, −120, and 120 degrees. This means that voltage
VA is in phase with V1mem; therefore, the impedance vector Z is
Fig. 25 Dynamic mho during fault conditions with varying sending source in phase with Zp.
angle: (a) −40°, (b) −20°, (c) 0°, (d) 20°. For an AG fault,
Distance-element overreach for high-resistance faults during VA
Z= (28)
heavy forward/reverse load conditions is a concern when I A + k0 ⋅ I R
applying distance elements. The adaptability of the positive-
sequence memory-polarized distance element minimizes V1mem
Zp = (29)
overreaching concerns [11]. I A + k0 ⋅ I R
 12

where IR represents the residual current, or sum of the phase

current (IA + IB + IC). Vectors Z and Zp overlap because they are X
in phase; therefore, the angle difference between Zp and dZ is 6
90 degrees when the measured impedance is on the self- Zr
polarized mho circle.
4
The graphical results shown in Fig. 26 for Tests 1 and 2 dZ
show that, in testing this way, the dynamic mho circle asserts at (a)
2
the same time as the self-polarized mho circle. With proper test Zp
voltages, the user does not see the effects of the dynamic mho Z R
circle. –6 –4 2 4

X –2

Zr
4 X
dZ
(a) 6
2 Zr
Zp
4
Z R
dZ
–6 –4 4 (b)
2
–2
Z R
–6 –4 –2 2 4
Zp

X
6 dZ
Zr Fig. 27 Uncommon test values for an AG fault; dynamic mho shifts
Z
4 (a) Test 3: Inside the static mho, (b) Test 4: Outside the static mho.
Zp
This same element behavior also applies to the phase
2 elements. From Table 2, we obtain the following:
(b)
R
VBC1mem =− j ⋅ VA1mem ⋅ 3
–6 –4 –2 2 4 6 VBC is calculated as
–2 VBC =VB − VC =− j ⋅ VA1 ⋅ 3
Both vectors are in phase using common test voltages;
–4
therefore, Vectors Z and Zp are also in phase, and the dynamic
mho circle does not affect testing. For a BC fault, we obtain the
–6 following:
VBC
Fig. 26 Common test values for an AG fault; the dynamic mho does not Z= (30)
interfere with testing when the angle of VA and V1mem are equal: (a) Test 1,
I B − IC
(b) Test 2.
VBC1mem
Test 3, shown in Fig. 27(a), shows an example of where
Zp = (31)
I B − IC
V1mem lags VA and the dynamic mho circle moves inside the
static mho. In this scenario, the element does not operate until C. Verification
the impedance enters the dynamic mho circle. Test 4, shown in Often, we can draw the dynamic and static mho elements on
Fig. 27(b), shows the opposite scenario, where V1mem leads VA, the same graph, but the dynamic mho element determines relay
causing a test impedance from the first quadrant to enter the operation. By performing a steady-state test, we can prove the
dynamic mho circle first and operate before reaching the static existence of the dynamic mho circle. We use the values shown
mho. in Table 5 to test a relay with positive-sequence memory
voltage. Settings are shown in Table 6.
 13

TABLE 5 V. PRACTICAL CONSIDERATIONS

VERIFICATION TEST QUANTITIES
A. Benefits of Positive-Sequence Memory Voltage
Test Parameter Applied Voltage Unit (Sec)
There are few settings related to the mho element that a user
VA 35∠40° V
must set. The most important of these include the reach point,
VB 40∠−100° V direction, and line angle or maximum torque angle (MTA).
VC 45∠160° V High-impedance fault coverage can be very limited for a self-
polarized mho element because the circle does not expand. We
V1 39.86∠13.3° V
can improve resistive coverage by increasing the reach point or
TABLE 6 varying the MTA, but this is unacceptable in most circumstances.
VERIFICATION RELAY SETTINGS Positive-sequence memory polarization provides excellent
Relay Setting Value Unit dynamic expansion with good open-pole security compared to
other polarizing types, such as self-polarization, cross-
Zone 1 reach 6.24 Ω (sec)
polarization, and hybrid polarization [1]. A larger dynamic
Line impedance 7.8 Ω (sec) expansion provides the benefit of additional resistive coverage,
Line angle 84 Degrees without sacrificing security or requiring any settings changes.
The expansion properties depend on several factors, including,
| k0 | 0.726 unitless
but not limited to, source and line parameters, load flow, fault
arg (k0) –3.69 Degrees type, fault resistance, mho implementation, and memory properties.
Positive-sequence memory voltage has the added benefit of
We set the current test angle to 0 degrees and vary the
providing proper polarization for faults on any phase and during
magnitude. Maintaining constant voltage and slowly increasing
single-pole tripping applications [1]. During single-pole open
the current ensures that the dynamic mho has minimal
intervals, the positive-sequence memory voltage polarization
expansion. The position of the static and dynamic mho circles
quantity remains stable enough for use with phase- and ground-
are shown in Fig. 28. The locus of impedance test points
mho elements.
Ztest_locus shows the impedance trajectory as the test current
The sequence network for a single-pole open on a line is
increases from 4.41 to 4.93 A (sec).
shown in Fig. 29. We assume each source contains a zero-
sequence source. The positive-sequence current is split between
X the parallel branches of negative- and zero-sequence
impedance. The positive-sequence voltage developed at the
relay during steady state is less than nominal three-phase
| IA | = 4.41 A
Zr voltage, but it is still sufficient for polarization.
4
Ztest_locus

2 | IA | = 4.93 A ES ER

ZS1 mZL1 (1 – m)ZL1 ZR1

R
–4 –2 2 4

Fig. 28 Uncommon test values interfere with mho element testing.

As the current magnitude increases, the dynamic circle

ZS2 mZL2 (1 – m)ZL2 ZR2
moves slightly in real-time. It is clear that the apparent
impedance test point does not intersect both mho circles at the
same point. Without a knowledge of the dynamic mho
expansion, a relay tester may expect ground mho operation
when | IA | = 4.41 A (sec) because this is the location at which
the apparent impedance intersects with the static mho ZS0 mZL0 (1 – m)ZL0 ZR0
characteristic. In reality, the current needs to increase until the
impedance enters the dynamic mho circle at | IA | = 4.93 A (sec) Fig. 29 Sequence networks with single-pole open.
before the element operates.
By contrast, ground quadrilateral elements provide user-
The dynamic mho circle rarely interferes with testing unless
settable resistance blinder thresholds for greater control of fault
the positive-sequence voltage angle differs from the faulted-
resistance coverage; however, these elements often use
phase voltage angle. The relay test voltages in the previous
negative- or zero-sequence current for polarization and must be
example could be manipulated to force the apparent impedance
disabled during single-pole open applications [5].
to intersect the dynamic mho circle before the static mho. In
this case, the ground mho element operates before the expected
trip current.
 14

B. Expansion and Security Relay R2 initially restrains for the three-phase reverse fault
Mho elements are inherently directional, as shown in Fig. 8 but may eventually operate as the dynamic mho shrinks back to
and Fig. 22. Forward faults expand the dynamic element back the self-polarized mho. A properly set directional element that
towards the source, and reverse faults shrink the element in the supervises the mho element prevents relay operation even if the
first quadrant. impedance plots inside the dynamic mho circle. The
As with the use of other elements, there are some instances progression of the dynamic characteristic is shown in
where using the mho element poses a security concern. An Fig. 32(a). If the fault is not three-phase, the dynamic mho will
example would be for ground-distance element overreach not completely shrink to the self-polarized mho, as described in
during phase-to-phase-to-ground faults [1]. We can largely Section III. In most cases, the reach point Zr exceeds the series
mitigate these security concerns by supervising the mho capacitor reactance Xc; however, if the opposite is true, then R2
characteristic with directional and fault-identification logic [4]. will operate during mho expansion and restrain during steady
In such a special case as a series capacitor application, we state, as shown in Fig. 32(b).
strongly recommend transient simulations to ensure proper
dynamic operation. The capacitive impedance in these X
scenarios can place more emphasis on dynamic expansion for
proper relay operation. For example, in Fig. 30 we apply a 4

three-phase fault close to relays R1 and R2, which have

positive-sequence memory polarization.
3
R2 R1
Percent
dZ Zr
ES ER Expansion
Xc (a)
2 100%
ZS1 ZL1 ZR1
50%
0%
3ph Zp
Z
Fig. 30 Relays R1 and R2 with bus-side PTs react to a three-phase fault
beyond a series capacitor on a local line.
1 ∙ ZS1
R
The impedance Z measured by R1 will be close to –2 –1 1 2
−90 degrees, as shown in Fig. 31, and the dynamic expansion
is crucial to proper relay operation. Several system parameters
affect relay operation. These parameters include the magnitude X
of the capacitive reactance, the extent of the dynamic
1
expansion, and filtering delays. The element initially operates
if | ZS1 | > | Xc | and eventually restrains as the dynamic circle Zp
Percent
1 ∙ ZS1
shrinks to the self-polarized element. Expansion
100%

dZ 80%
X
(b) 60%
3 Z
40%
Zr
20%
2 Zr
0%
0.2

1 dZ

R
–0.4 –0.2 0.2 0.4
R
–2 –1 1 2 Fig. 32 Impedance diagram for Relay R2 in response to a reverse three-
Z phase fault on the opposite side of the series capacitor: (a) Case I: | Zr | > | Xc |,
Zp (b) Case II: | Zr | < | Xc |.

These examples highlight the importance of transient testing

with series-compensated lines. Real-world cases can be
Fig. 31 Impedance diagram for Relay R1 in response to a forward three-
significantly more involved when we consider other factors
phase fault on the opposite side of the series capacitor. such as parallel-compensated lines, bus versus line PTs,
midline versus end compensation, and subharmonic
transients [8].
 15

VI. CONCLUSION VIII. BIOGRAPHY

Dynamic mho elements with positive-sequence memory Donald D. Fentie received his B.Sc. degrees in Electrical Engineering and
Computer Science in 2006 and his M.Sc. degree in Electrical Engineering in
polarization have large expansion and excellent single-pole 2010 from the University of Saskatchewan. Mr. Fentie worked as a protection
open security. This expansion provides additional resistive engineer and field engineer for Magna Electric Corporation in Saskatoon,
coverage during faults, directional security, and protection for Saskatchewan from 2006 to 2010. He joined Schweitzer Engineering
Laboratories, Inc. (SEL) in 2010 and is currently an application protection
close-in faults. The time-varying nature of the element can be engineer in Calgary, Alberta. He is an IEEE member and is registered as a
visualized with vectors on the impedance plane. Professional Engineer (P.Eng.) in the province of Alberta.
The dynamic mho behavior is affected by system
parameters, load flow, fault type, and the relay algorithm. Real-
time transient simulations are recommended for applications
such as series-compensated lines, where the dynamic mho
element is expected to play a crucial role in relay operation.
A thorough understanding of the distance-element concepts
presented in this paper is intended to help in the following
ways:
• Relay testers can avoid misinterpreting results when
performing transient and steady-state tests.
• Disturbance-analysis teams can assess the performance of
a mho element based on changing system conditions.
• Users creating settings can better understand system
conditions that adversely affect mho performance.

VII. REFERENCES
[1] E. O. Schweitzer, III and J. Roberts, “Distance Relay Element Design,”
Proceedings of the 46th Annual Conference for Protective Relay
Engineers, College Station, TX, April 1993.
[2] E. O. Schweitzer, III, “New Developments in Distance Relay Polarization
and Fault Type Selection,” presented at the 16th Annual Western
Protective Relay Conference, Spokane, WA, October 1989.
[3] F. Calero, “Distance Elements: Linking Theory With Testing,” 62nd
Annual Conference for Protective Relay Engineers, Austin, TX,
April 2009, pp. 333–352.
[4] D. Costello, “Determining the Faulted Phase,” 63rd Annual Conference
for Protective Relay Engineers, College Station, TX, April 2010,
pp. 1–20.
[5] H. Altuve Ferrer and E. O. Schweitzer, III, “Modern Solutions for
Protection, Control, and Monitoring of Electric Power Systems,”
Schweitzer Engineering Laboratories, Inc., Pullman, WA, 2010.
[6] S. Zocholl, “Three-Phase Circuit Analysis and the Mysterious k0 Factor,”
22nd Annual Western Protective Relay Conference, Spokane, WA,
October 1995.
[7] L. M. Wedepohl, “Polarized Mho Distance Relay,” Proceedings of IEE,
Volume 112, No. 3, March 1965.
[8] H. Altuve, J. Mooney, G. Alexander, “Advances in Series-Compensated
Line Protection,” 35th Annual Western Protective Relay Conference,
Spokane, WA, October 2008.
[9] J. Zydanowicz, “Application of the Idea of Steady-State Impedance and
Admittance to the Construction of Diagrams Intended for the Analysis of
the Operation of Distance and Directional Relays and Protective
Schemes,” Conférence Internationale des Grands Reseaux Électriques à
Haute Tension, Paris, France, June 1960.
[10] L. P. Cavero, “Analysis of Complex Distance-Relay Characteristics
Taking Load Into Account,” IEE Conf. Publ. 185, 1980, pp. 192–194. Previously presented at the 2016 Texas A&M
[11] J. Roberts, A. Guzman, and E. O. Schweitzer III, “Z=V/I Does Not Make Conference for Protective Relay Engineers.
a Distance Relay,” presented at the 20th Annual Western Protective Relay © 2016, 2017, 2023 IEEE – All rights reserved.
Conference, Spokane, WA, October 1993. 20230911 • TP6719-01