Input Source Error Concerns for

Protective Relays

David Angell
Idaho Power

Daqing Hou
Schweitzer Engineering Laboratories, Inc.

Presented at the
61st Annual Georgia Tech Protective Relaying Conference
Atlanta, Georgia
May 2–4, 2007

Previously presented at the

60th Annual Conference for Protective Relay Engineers, March 2007

Originally presented at the

33rd Annual Western Protective Relay Conference, October 2006
 1

Input Source Error Concerns for

Protective Relays
David Angell, Idaho Power
Daqing Hou, Schweitzer Engineering Laboratories, Inc.

Abstract—Capacitive Coupled Voltage Transformers circuit and equivalent circuit are shown in Fig. 1 and Fig. 2,
(CCVTs) and Bushing Potential Devices (BPDs) have supplied respectively. The equivalent circuit includes stray capacitance
high voltage relay potential circuits for many years. The and winding resistance of the compensating reactor and also
transient performance of a CCVT has been analyzed, and
solutions for their transient performance have been incorporated
stray capacitance, winding resistance, and leakage inductance
in relaying systems since the 1970s. Little analysis has been of the step-down transformer.
performed on BPD transient performance. A BPD has been
Line Voltage
tested and characterized to determine the parameters that affect
the output voltage during system fault conditions.
Current Transformers (CTs) that saturate during system C1
faults produce a nonsinusoid output that, when filtered, presents GAP RR
a reduced magnitude and phase shifted current to protective
relays. The response of distance relays for various levels of CT
saturation are analyzed and presented. L TR
GAP CF LF
B
Finally, utility experience with relays sourced by CCVTs and C2
U
R
BPDs with the intent of validating the models and relay setting D
adjustments for input source errors are presented. CA RA RF E
N

I. INTRODUCTION Step Down

Electromechanical relays are reasonably tolerant of voltage Voltage
source errors due to their inherent signal filtering and slower Transformer
response. The filtering and response of solid-state and Fig. 1 CCVT circuit
numerical relays have been tuned to provide fast response to
CC
fundamental frequency values. This fast response has enabled
them to respond during the transient period of the voltage
sources. CTs that were installed many years ago may be
subjected to higher fault currents than originally contemplated.
These CTs may saturate during the system faults, resulting in Ce LC RC LP RP B
LF U
an erroneous signal being applied to the relay. Protection CF R
CP D
engineers may specify voltage and current source suitable for E
RF
relaying performance on new construction. However, there are N

many relay retro-fit situations where the project budget will

not tolerate replacement of voltage and current transformers. Fig. 2 CCVT equivalent circuit
This paper will review CCVT modeling, present a model
for BPDs, analyze distance relay performance during CT B. CCVT Voltage Transient
saturation, and provide utility experience with CCVT and For a few cycles following a voltage change, the output of
BPD performance and transient error mitigation. a CCVT does not match the input. This output error is known
as a subsidence voltage. The typical demonstration of
II. CAPACITIVE COUPLED VOLTAGE TRANSFORMER subsidence voltage is to plot the voltage response for a short
circuit occurring at voltage peak and zero as shown in Fig. 3
A. CCVT Construction and Fig. 4, respectively. These transient voltage responses are
A CCVT is a device that makes use of a capacitive voltage from the equivalent active CCVT circuit shown in Fig. 2.
divider to reduce the primary voltage to a medium voltage
level, e.g. 15 kV. The medium voltage is applied to a voltage
transformer that delivers the common secondary 115 and 66
voltages.
The other components associated with the CCVT are a
compensating reactor and ferroresonance suppression circuit.
A typical CCVT with an active ferroresonant suppressing
 2

100
TRANSIENT RESPONSE: SOLID-Ratio Voltage, DASH-CCVT Output The subsidence voltage causes a measured voltage
magnitude reduction and phase angle shift as shown in the
80
polar plot, Fig. 6 and Fig. 7, for peak and zero voltage fault
60 initiation. The numbers marked in the plots are the sequence
40 that the fault voltage goes through after the fault initiation.
20
120 60
volt

-20

150 30
-40

-60

-80
4 2
3
-100
-2 -1 0 1 2 3 4 5 5
180 0
cycle

Fig. 3 CCVT peak voltage subsidence transient

TRANSIENT RESPONSE: SOLID-Ratio Voltage, DASH-CCVT Output

100
I
80
210 330

60

40
240 300
20
270
volt

0
Fig. 6 CCVT voltage peak subsidence transient fundamental voltage
-20

120 60
-40

-60

-80
150 5 30
4
-100
-2 -1 0 1 2 3 4 5
cycle

Fig. 4 CCVT zero voltage subsidence transient

6
C. Explanation of Transient 180 0

For a voltage peak transient, the capacitor is fully charged 3

and is discharged with an associated RLC time constant into 2

the short circuit. The resulting equivalent is an underdamped
circuit that will result in a transient voltage that oscillates at a 210 330

frequency greater than the fundamental as shown in Fig. 3. For

a voltage zero transient, the transformer and compensating
I
reactor are at peak flux. This stored flux is released through
240 300
the capacitors to the short circuit with a RLC time constant.
This results in an overshoot voltage that rings down to zero in 270

a few cycles as shown in Fig. 4. A simplified equivalent Fig. 7 CCVT voltage zero subsidence transient fundamental voltage
circuit is shown in Fig. 5. Notice how the voltage magnitude is severely reduced and
significantly out of phase with the actual fault voltage, points
Ls Ce Lc 5 and 6, respectively in Fig. 6 and Fig. 7.

Rb

Vs I Vo
Lb

Fig. 5 Simplified CCVT equivalent circuit for a terminal short circuit

 3

6
IMPEDANCE TRAJECTORY: SOLID-Ratio Voltage, DASH-CCVT Transient E. Utility CCVT Performance
IMPEDANCE TRAJECTORY: SOLID-Ratio Voltage, DASH-CCVT Transient

Two examples of CCVT performance are provided. In both

100

90

cases the numerical relays are applied to the secondary,

80

5 70

resulting in a low secondary burden. The first example is a

60

X-ohm
50

4
failed circuit switcher contact on a line reactor that is selected
40

30

to open for voltage control. The Phase B contact remains

20

3 10

closed as the horizontal air break is opened toward Phase C

0
-20 0 20 40 60 80 100 120 140 160 180 200
R-ohm
X-ohm

2
creating a B-C fault. A remote Zone 1 relay, set to protect a
1
41-mile line with a SIR of 1.45, trips for this event. The
CCVT sourcing the protective relay has an active
0 ferroresonance suppression circuit. The resulting transient
impedance lies within the Zone 1 reach due to the subsidence
-1
transient as shown in Fig. 11 and Fig. 12.
-2
-2 -1.5 -1 -0.5 0 0.5 1 BC-Phase Voltage and Current
R-ohm 150

100
Fig. 8 CCVT output impedance trajectory of voltage-peak fault
50

volt
IMPEDANCE TRAJECTORY: SOLID-Ratio Voltage, DASH-CCVT Transient 0
10 IMPEDANCE TRAJECTORY: SOLID-Ratio Voltage, DASH-CCVT Transient
50
-50
40

9 30 -100
20

8 10 -150
0 2 4 6 8 10 12
X-ohm

7 -10

-20

-30 15
6
-40

-50
0 20 40 60 80 100 120 140 160 180 200 10
5 R-ohm
X-ohm

5
amp

4
0
3
-5
2
-10
0 2 4 6 8 10 12
1 cycle

0
Fig. 10 B-C phase voltage and current for B-C fault
-1
-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4
BC-Phase Apparent Impedance
R-ohm 6

Fig. 9 CCVT output impedance trajectory of voltage-zero fault

5

This voltage error causes the relay measured impedance

trajectory to present a negative-impedance reach before 4

traversing through zero to the actual fault impedance as shown

in Fig. 8 and Fig. 9, where the main graph shows the enlarged 3
Im-ohm

version of the entire impedance trajectory plotted in a small

graph of the same figure. These examples present cases where 2

the Zone 1 would have to be eliminated because it could not

be set short enough to avoid misoperation. 1

D. Parameters That Affect CCVT Transient Performance 0

TABLE I
PARAMETERS THAT AFFECT CCVT TRANSIENT PERFORMANCE -1
-3 -2 -1 0 1 2 3 4 5
Re-ohm
Parameter Small Large
Transient Transient Fig. 11 Zone 1 overreach for B-C fault

CCVT Capacitance High Low The second case occurred when a breaker was closed into a
three-phase fault. A remote Zone 1 relay, set to protect a
Ferroresonance Suppression Passive Active
Circuit 13-mile line with a SIR of 29, trips for this event. The
resulting transient impedance lies within the Zone 1 reach due
Transformer Ratio High Low
to the subsidence transient as shown in Fig. 12 and Fig. 13.
Burden Resistive Inductive
Bus Voltage Dip* Small Large
* The transient error is proportional to the
change in bus voltage due to a fault.
 4

100
Three-Phase Voltages and Currents: SOLID-A,DASH-B,DOT-C The secondary of the BPD contains a coupling coil, phase
angle adjustment capacitors, and power factor adjustment
50
capacitors. The coupling coil and phase angle adjustment
capacitors provide for aligning the secondary potential with
volt

the primary. This adjustment is required because the primary

-50
capacitance will vary based on the primary operating voltage.
-100
0 2 4 6 8 10 12 The power factor adjustment capacitors are to compensate the
secondary burden to near unity power factor to minimize
10 phase shift and voltage drop due to the burden. BPDs are
5 notoriously difficult to adjust for electromechanical relaying
0 burdens.
amp

-5
BPDs are designed to withstand overvoltage up to rated
-10
line-to-line voltage that result when neutral voltage shift
-15
occurs. Therefore, a ferroresonance suppression circuit is not
0 2 4 6
cycle
8 10 12
required. A typical BPD and equivalent circuit are shown in
Fig. 12 Voltages and currents for three-phase fault Fig. 14 and Fig. 15, respectively. Refer to the Appendix for
Three-Phase Apparent Impedance: SOLID-A,DASH-B,DOT-C
the model parameters.
0.8

0.7
C1 SW
0.6

Secondary
0.5 Cpa Cpf 115

Coupling
Primary
C2 C3 66.4

Coil
0.4
Im-ohm

SW
0.3

0.2
Fig. 14 GE KA-105 BPD circuit
0.1
L3 R3
0
L2 R2 m I2 + I3
-0.1
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 Lb
Re-ohm

Cpa
Fig. 13 Zone 1 overreach for three-phase fault Ce L1 R1 V3 Cpf Vb
1-m
In the first case, the Zone 1 reach was reduced to eliminate I1 Rb
the transient overreach. In the second case, a 1.75-cycle time Vp V1 V2
I2 I3
delay was added to eliminate the overreach. N1 N2 N3
F. Recommended CCVT Specification Primary Coupling Secondary Burden

In order to minimize the CCVT subsidence transient, use Fig. 15 GE KA-105 BPD equivalent circuit
the following criteria: B. BPD Voltage Transient
• High capacitance, e.g. 10 nano-fared
A BPD produces a similar response to a change in voltage
• Large transformer ratio, e.g. 15 kV / 115 / 66 V to that of a CCVT. However, because the BPD employs a low
• Passive ferroresonance circuit step-down transformer ratio, the BPD response is more
• Resistive CCVT burden dependent on the secondary burden as shown in Fig. 16
through Fig. 19. A high inductive burden (low power factor)
III. BUSHING POTENTIAL DEVICE significantly increases the amount and duration of the transient
A. BPD Construction error as shown in Fig. 18 and Fig. 19.
BPDs are manufactured to match a particular bushing type,
e.g. GE “U” or “F” type. The BPD may be applied on a
bushing type ranging from 69 to 230 kV. The construction of
a BPD is based on tapping a point within the degrading
insulation of a bushing. This makes use of the capacitive
effect of a breaker bushing to produce a capacitive voltage
divider. The BPD reduces the primary voltage to a medium
voltage level, e.g. 4 kV, which is applied to a transformer to
produce the secondary potentials of 115 and 66 volts.
 5

Voltage Peak: SOLID-ratio voltage, DASH-secondary Voltage Peak: SOLID-ratio voltage, DASH-secondary
1 1

0.5 0.5

volt
volt

0 0

-0.5
-0.5

-1
-1 -0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06

Voltage Zero, VA=30.0; pf=0.7

Voltage Zero, VA= 5.0; pf=1.0 1
1

0.5
0.5

volt
0
volt

-0.5
-0.5

-1
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06
-1
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 seconds
seconds

Fig. 16 BPD response for burden of VA=5 and pf=1.0 with exact pf Fig. 19 BPD response for burden of VA=30 and pf=0.7 with exact pf
compensation compensation

1
Voltage Peak: SOLID-ratio voltage, DASH-secondary C. Parameters That Affect BPD Transient Performance
• Burden: Inductive burdens result in worst transient
0.5
performance.
• Transformer Ratio: A low step-down transformer ratio
volt

-0.5 will result in a larger burden effect and produce

greater transient error.
• Power Factor Adjustment: Compensating for any
-1
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06

Voltage Zero, VA=30.0; pf=1.0 lagging power factor burden with capacitor Cpf will
improve the transient performance.
1

0.5 • Bus Voltage Dip: The transient error is proportional to

the change in bus voltage caused by a fault. Therefore,
volt

high source-to-line-impedance ratios result in these

large voltage changes and transient error at Zone 1
-0.5

-1
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06
boundary faults.
seconds
D. Utility BPD Performance
Fig. 17 BPD response for burden of VA=30 and pf=1.0 with exact pf
compensation Two examples are provided demonstrating BPD
performance. The first case involves a relaying system
Voltage Peak: SOLID-ratio voltage, DASH-secondary
1 consisting of both numerical and electromechanical relays
0.5
resulting in a high burden, 10 ohms at 65% power factor. The
protected line is 40 miles long with an SIR of 5.8. A ground
volt

0
fault occurs 64% from a station. The large and slow decaying
-0.5 BPD transient results in an underreach for this fault as shown
-1
in Fig. 20 and Fig. 21.
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06

Voltage Zero, VA= 5.0; pf=0.7

1

0.5
volt

-0.5

-1
-0.02 -0.01 0 0.01 0.02 0.03 0.04 0.05 0.06
seconds

Fig. 18 BPD response for burden of VA=5 and pf=0.7 with exact pf
compensation
 6

C-Phase Voltage and Current CA-Phase Apparent Impedance

100 2.5

50
2
volt

-50 1.5

Im-ohm
-100
0 2 4 6 8 10 12
1

10
0.5

5
amp

0 0

-0.4 -0.2 0 0.2 0.4 0.6 0.8

-5 Re-ohm

-10
0 2 4 6 8 10 12 Fig. 23 C-A phase apparent impedance
cycle
E. Mitigation of CCVT and BPD Subsidence Transient
Fig. 20 Voltage and current for a C-phase fault at 64% of line length
The characteristic of the CCVT and BPD subsidence
18
C-Phase Apparent Impedance
transient is a reduction in fundamental voltage for a few
cycles. This will result in a Zone 1 distance relay overreach
16
for remote terminal faults. The preferred mitigation to apply
14
depends on the fault clearing requirements of the transmission
12 system. If the system can tolerate a few cycles for fault
10
detection, then time delay the Zone 1 output by up to two
cycles. However, for the system with stringent fault clearing
Im-ohm

8
times, the Zone 1 relay reach should be reduced based on the
6 chart shown in Fig. 24. Fig. 24 is obtained using typical
4
passive and active CCVTs with a resistive burden of 5 kΩ,
representative of the load of several modern microprocessor
2
relays and meters.
0 1

-4 -3 -2 -1 0 1 2 3 4 5 6
0.9
Re-ohm

0.8
Fig. 21 Apparent impedance of a C-phase fault at 64% of line length
max. zone 1 reach setting (pu)

0.7
The second case involves all numerical relays presenting a
Passive CCVT

low burden to the BPD. A remote Zone 1 relay, set to protect a

0.6

3-mile line with a SIR of 1.87, trips when a clamp fails, 0.5

allowing Phase A to fall into Phase C on the station dead-end 0.4

structure one station beyond the remote end bus. The resulting 0.3 Active CCVT

transient impedance lies within the Zone 1 reach due to the 0.2

subsidence transient as shown in Fig. 22 and Fig. 23. 0.1

CA-Phase Voltage and Current 0

150 0 5 10 15 20 25 30
SIR
100

50
Fig. 24 Distance element performance as function of source impedance ratio

In some microprocessor relays, dedicated logic has been

volt

-50
developed to deal with the transient overreach concerns from
-100

-150
CCVT transients. The system impedance ratio (SIR) is
0 2 4 6 8 10 12
estimated from the fault voltage and current, when this value
100
is high and causes a concern of overreach, a time delay is
added to the instantaneous tripping distance elements. The
50
fault impedance is then closely monitored to detect any CCVT
transient signatures. Therefore, on the detection of a high SIR,
amp

-50
the relay applies the Zone 1 delay and monitors the voltage
transient. If the transient signature is small and does not
-100
0 2 4 6
cycle
8 10 12 indicate overreach, the time delay is quickly removed to allow
a quicker operation of the elements. The logic adapts to the
Fig. 22 C-A phase voltage and current quality of the CCVT used and only adds time delay when
necessary.
 7

IV. CURRENT TRANSFORMER 20

Currents: SOLID-CT Output, DASH-Ratio

A. CT Saturation 10

CT saturation results in a reduction in fundamental 0

amp
magnitude and a leading phase shift. This error will cause -10

distance relays to under reach a remote fault as shown in -20

Fig. 25 through Fig. 30. -30

-3 -2 -1 0 1 2 3 4 5 6 7

Currents: SOLID-CT Output, DASH-Ratio

20 Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10
10
8
0
amp

pu
-10
4

-20
2

-30
-3 -2 -1 0 1 2 3 4 5 6 7 0
-3 -2 -1 0 1 2 3 4 5 6 7
cycle
Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10
Fig. 28 CT output and distance measurement with 6.3 ms to saturation
8
Currents: SOLID-CT Output, DASH-Ratio
6 20
pu

4 10

2 0

amp
0 -10
-3 -2 -1 0 1 2 3 4 5 6 7
cycle
-20

Fig. 25 CT output and distance measurement without saturation -30

-3 -2 -1 0 1 2 3 4 5 6 7

Currents: SOLID-CT Output, DASH-Ratio

20 Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10

10
8

0
amp

6
pu

-10
4
-20
2

-30
-3 -2 -1 0 1 2 3 4 5 6 7 0
-3 -2 -1 0 1 2 3 4 5 6 7
cycle
Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10
Fig. 29 CT output and distance measurement with 4.5 ms to saturation
8
Currents: SOLID-CT Output, DASH-Ratio
6 20
pu

4 10

2 0
amp

0 -10
-3 -2 -1 0 1 2 3 4 5 6 7
cycle
-20

Fig. 26 CT output and distance measurement with minimal saturation -30

-3 -2 -1 0 1 2 3 4 5 6 7

Currents: SOLID-CT Output, DASH-Ratio

20 Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10
10
8
0
amp

6
pu

-10
4
-20
2
-30
-3 -2 -1 0 1 2 3 4 5 6 7 0
-3 -2 -1 0 1 2 3 4 5 6 7
cycle
Distance Calculations: SOLID-Using CT Output, DASH-Using Ratio Current
10
Fig. 30 CT output and distance measurement with 3.3 ms to saturation
8

6
pu

0
-3 -2 -1 0 1 2 3 4 5 6 7
cycle

Fig. 27 CT output and distance measurement with 7.7 ms to saturation

 8

The time to saturate determines the reduction in reach. TABLE III

Table II summarizes the reach reduction based on the time to Parameter Description Value
saturate.
Ce Equivalent Capacitance (C1+C2) 0.0059 µF
TABLE II
DISTANCE RELAY REACH DURING CT SATURATION L1 Transformer Primary Leakage Inductance 0.263 mH
R1 Transformer Primary Winding Resistance 0.22 Ω
Time to Saturate Reach
(milliseconds) (% at 1.5 cycles) Cpa Phase Angle Adjustment Capacitance 12.9 µF

No saturation in first cycle 100 L2 Coupling Coil Leakage Inductance 0.298 mH

7.7 50 R2 Coupling Coil Winding Resistance 0.077 Ω

6.3 33 Transformer Secondary Leakage

L3 3.836 mH
Inductance
4.5 25
Transformer Secondary Winding
3.3 17 R3 0.086 Ω
Resistance
Adjusted
B. Conclusion Cpf Power Factor Adjustment Capacitance
for burden
Utility experience with CCVT transients has demonstrated
that the CCVT model is accurate and can be used to determine
VI. REFERENCES
mitigation methods. Both utility examples of reach reduction
[1] M. Kezunovic, C. W. Fromen, and S. L. Nilsson, “Digital Models of
and time delay have produced satisfactory results for dealing Coupling Capacitor Voltage Transformers for Protective Relay
with existing CCVT sources. Additionally, proper CCVT Transient Studies,” IEEE Transactions on Power Delivery, Vol. 7,
specifications will result in better performing CCVTs for new No. 4, October 1992.
installations. [2] A. Sweetana, “Transient Response Characteristics of Capacitive
A model for BPDs has been developed and validated with Potential Devices,” IEEE Transactions on Power Apparatus and
Systems, Vol. 90, No. 5, September/October 1971.
captured events. The analysis of BPDs has determined that
[3] D. Hou and J. Roberts, “Capacitive Voltage Transformers: Transient
these devices do produce a subsidence transient. This transient Overreach Concerns and Solutions for Distance Relaying,” presented at
is dependent on the burden attached to the device. Ultimately, the 22nd Annual Western Protective Relay Conference, Spokane, WA,
BPDs are suitable for protective relaying applications using October 24–26, 1995.
numerical relays and meters that present a low, near unity
power factor burden. VII. BIOGRAPHIES
CT saturation will cause distance relay underreach. Where David Angell graduated from the University of Idaho with a bachelors degree
CT saturation has been identified and there is no ability to in electrical engineering in 1984 followed with a masters degree in 1986. He
has 21 years of service with Idaho Power and two years with Bonneville
mitigate the saturation with CT settings (e.g. changing the Power Administration. David has 20 years experience in power system
ratio or reducing the burden) or replacement, then the reach of protection and communications. He has presented several papers on protection
the Zone 2 relays should be adjusted for this underreach. topics for conferences and university lectures. Presently, David is the manager
of Planning and Load Research for Idaho Power and an adjunct professor at
Boise State University.
V. APPENDIX
This appendix shows the equivalent circuit and parameters Daqing Hou received BS and MS degrees in electrical engineering at the
Northeast University, China, in 1981 and 1984, respectively. He received his
used to model a GE KA-105 Bushing Potential Device. Ph.D. in electrical and computer engineering at Washington State University
in 1991. Since 1990, he has been with Schweitzer Engineering Laboratories,
L3 R3
Inc., Pullman, Washington, USA, where he has held numerous positions
L2 R2 m I2 + I3
including development engineer, application engineer, and R&D manager. He
is currently a principal research engineer. His work includes system modeling,
Lb simulation, and signal processing for power systems and digital protective
relays. His research interests include multivariable linear systems, system
Cpa
Ce L1 R1 V3 Cpf Vb identification, and signal processing. He holds multiple patents and has
1-m
authored or co-authored many technical papers. He is a Senior Member of
I1 Rb IEEE.
Vp V1 V2
I2 I3
N1 N2 N3
Primary Coupling Secondary Burden

Fig. 31 BPD equivalent circuit

Previously presented at the 2007 Texas A&M

Conference for Protective Relay Engineers.
© 2007 IEEE – All rights reserved.
20070209 • TP6269-01