ASV KYV

ES1 ES2
SR1 SR2 SR3 SR4
L1 L2

Line Differential Protection Scheme Modelling for
Underground 420 kV Cable Systems
EMTDC/PSCAD Relays Modelling

Michal Sztykiel, Claus Leth Bak
Department of Energy Technology
Aalborg University
Aalborg, Denmark
clb@iet.aau.dk
Wojciech Wiechowski, Sebastian Dollerup
Protection & Control Systems
Energinet.dk
Fredericia, Denmark
sdo@energinet.dk


Abstract— Based on the analysis of a specific relay model and an
HVAC (High Voltage Alternating Current) cable system, a new
approach to EMTDC/PSCAD modelling of protective relays is
presented. Such approach allows to create complex and accurate
relay models derived from the original algorithms. Relay models
can be applied with various systems, allowing to obtain the most
optimal configuration of the protective relaying. The present
paper describes modelling methodology on the basis of Siemens
SIPROTEC 4 7SD522/610. Relay model was verified
experimentally with its real equivalent by both EMTP-simulated
and real world generated current signals connected to the relay.
Keywords— line differential protection; XLPE HVAC cable;
EMTDC/PSCAD relay model; SIPROTEC 4 7SD522; 7SD610
I. INTRODUCTION
Relay computer modelling is an important issue for
establishing proper protection scheme for the specified system.
Nowadays, it is difficult to obtain accurate relay computer
model, since relay manufacturers offer their products with a
variety of algorithms and features that may significantly
change operation of relays under specific conditions and states.
For this purpose, essential study over new approach for
creating relay models in EMTDC/PSCAD is given. The
methodology relies on obtained relay’s technical specification
(given by relays manufacturer), so that unique features and
algorithms - characteristic for each relay type, can be
developed. As a result, this would give complex relay model
narrowed and useful only for specified type of relay. In
compensation, relay computer model would be very accurate
(mainly in terms of sensitivity and operating speed) with easy
and user-friendly configuration panel, which is programmed
with the same parameter values as in real devices.
Established relay models would easily allow to perform
simulations of chosen study cases and examine possibilities of
unwanted tripping that might occur (e.g. due to transient power
electronics switching, overvoltages, external faults,
energization states, etc.).
This paper presents such analysis for Siemens line
differential relays SIPROTEC 4 SD522/610, as these relays are
planned to protect HVAC underground transmission cable
system built in Denmark in years 2012-2014.
Large capacitance of underground cables in comparison to
overhead lines brings original issues for the differential
protection scheme to consider, as both steady and transient
states have to be deeply analyzed. For steady state, charging
current is the factor that mostly affects relays function. For
transient states, relays may be affected by inrush currents that
occur due to shunt reactors switching operations (necessary for
reactive power compensation).
In order to properly reflect cable system’s influence over
relay’s current signals in mentioned states, cable system is
modelled with the usage of EMTDC/PSCAD software, as it
provides satisfactory accuracy for both steady and transient
analysis.
When both cable system and protection scheme models are
completed, relay model’s accuracy can be finally verified
through experimental testing. Having identical parameter
setting both for relay model and real device, sensitivity and
operating speed are compared thus showing high accuracy of
the relay model.
Results from experimental analysis prove that presented
approach for relay modelling can be successfully adapted for
specific relays with original algorithms and features.
II. PROTECTED CABLE SYSTEM - BACKGROUND
A. Cable System - Description
The single phase diagram of total cable system is shown on
Fig. 1.
Figure 1. Schematic Representation of the Underground 420 kV Cable
System.

Fault Type Control
-------------------------------
1 = NoFault
2 = Phase Ato Ground
3 = Phase Bto Ground
4 = Phase CtoGround
5 = Phase AB toGround
6 = Phase ACto Ground
7 = Phase BCto Ground
8 = Phase ABCtoGround
9 = Phase AB
10 =PhaseABC
C1XLPE
S1
C2
S2
C3
S3
A
B
C
C1XLPE10
S1
C2
S2
C3
S3
C1XLPE2
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1 XLPE2
S1
C2
S2
C3
S3
C1XLPE3
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1 XLPE3
S1
C2
S2
C3
S3
C1XLPE4
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1 XLPE4
S1
C2
S2
C3
S3
C1XLPE5
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1 XLPE5
S1
C2
S2
C3
S3
A
B
C
C1XLPE6
S1
C2
S2
C3
S3
A
B
C
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1XLPE6
S1
C2
S2
C3
S3
C1 XLPE7
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
XLPE6 C
C1XLPE7
S1
C2
S2
C3
S3
C1 XLPE8
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1XLPE8
S1
C2
S2
C3
S3
C1 XLPE9
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1XLPE9
S1
C2
S2
C3
S3
C1 XLPE10
S1
C2
S2
C3
S3
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
C1 XLPE
S1
C2
S2
C3
S3
5.61 [H]2.858 [ohm]
5.61 [H]2.858 [ohm]
5.61 [H]2.858 [ohm]
KYVr
5.61 [H]2.858 [ohm]
5.61 [H]2.858 [ohm]
5.61 [H]2.858 [ohm]
A B C
ASVr
4.01 [H]2.159 [ohm]
4.01 [H]2.159 [ohm]
4.01 [H]2.159 [ohm]
4.01 [H]2.159 [ohm]
4.01 [H]2.159 [ohm]
4.01 [H]2.159 [ohm]
C B A
TOR2r
KYVr
ASVr
TOR2r
TOR1r
Timed Breaker Logic Open@t0
Timed Breaker Logic Open@t0
Timed Breaker Logic Open@t0
Timed Breaker Logic Open@t0
Ip1
Ip1
A
B
C
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
0.001 [ohm]
XLPE C XLPE2 C XLPE3 C XLPE4 C XLPE5 C
XLPE7 C XLPE8 C XLPE9 C XLPE10 C
Ip2
C B A
TOR1r
EN_1
EN_1
Timed Breaker Logic Closed@t0
Timed Breaker Logic Closed@t0 EN_2
EN_2
Ip2
Timed Fault Logic
A B C
FAULTS C
B
A
Timed Fault Logic
A B
V
Ph C
400.0 [kV], 50.0 [Hz]
10592 [MVA]
Z1 = 0.829 [ohm] + j16.6 [ohm]
A B
V
Ph C
400.0 [kV], 50.0 [Hz]
10477 [MVA]
Z1 = 0.839 [ohm] + j16.78 [ohm]
P_KYV
Q_KYV
P_ASV
Q_ASV
KYV
180
-180
PhaseAngle_1
0
degrees
PhaseAngle_1
0
500
0
Voltage_1
400
kV
Voltage_1
400
P_KYV
944
Q_KYV
82.6045
Main: Phasecurrents
0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200 ...
...
...
-1.00
-0.80
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
kA
Ia1 Ia2
-0.80
-0.60
-0.40
-0.20
0.00
0.20
0.40
0.60
0.80
kA
Ia1 Ia2
-10.0 -8.0
-6.0 -4.0
-2.0 0.0 2.0 4.0 6.0
8.0 10.0
kA
Ia1 Ia2
Faults : Controls
10
9
8
7
6
5
4
3
2
1
External Fault
0
Sources Control Set Faults Control Set Faults Control Set
Description
Line Circuit Breakers
Control Set
Shunt Reactors Circuit Breakers
Control Set
Phase Currents Waveforms - both line ends
EXTERNAL FAULT INTERNAL FAULT
KYV source
ASV source
KYV shunt reactor
TOR1 shunt reactor TOR2 shunt reactor
ASV shunt reactor
ASV
180
-180
PhaseAngle_2
0
degrees
PhaseAngle_2
0
500
0
Voltage_2
400
kV
Voltage_2
400
P_ASV
1304.54
Q_ASV
444.498
Faults : Controls
10
9
8
7
6
5
4
3
2
1
Internal Fault
1

ASV KYV
Protected Zone

FO
DR DR
CT CT

The system consists of the following components:
1) XLPE HVAC Underground Cable Sections (L1–L2):
Selected cable is made of three aluminium single-core cables
buried underground on the depth of 1,3 m and laid in a flat
formation within 300 mm from each other. Total cable is
divided into two sections of lengths accordingly 28 km and
29,5 km. Metallic screens of each cable section are cross-
bonded approximately each 2 km, and earthed each 6 km.
Detailed information about cable structure is presented on
table I.
TABLE I. TECHNICAL DATA OF XLPE UNDERGROUND 420 KV CABLE
Description Value
Cross-section of conductor (mm
2
) 1600
Diameter of conductor (mm) 52
Insulation thickness (mm) 27,0
Diameter over insulation (mm) 110,0
Cross-section of screen (mm
2
) 185
Outer diameter of cable (mm) 127,0
Capacitance (µF/km) 0,21
Inductance (mH/km) 0,50
Charging current per phase (A/km) 14,9

2) Shunt Reactor Banks (SR1–SR4): For reactive power
compensation, four switchable shunt reactors are installed;
each on ASV, KYV bus bars with reactive power of 100
MVARs and two between cable sections with reactive power
of 140 MVARs.
3) Supply Sources (ES1–ES2): Power system on both sides
of the cable is modelled by ES1 and ES2 sources that are
Thevenin equivalents consisting of voltage sources and its
short-circuit impedances. Parameter values are listed on table
II.
TABLE II. TECHNICAL DATA OF CABLE’S SUPPLY SOURCES
Supply Source
Voltage
(kV)
Short-circuit impedance
(Ω)
ES1 420 0.829 + j16.60
ES2 420 0.839 + j16.78

B. EMTDC/PSCAD Model of Cable System
In Fig. 2, described cable system is modelled in
EMTDC/PSCAD software by frequency dependent (phase)
model, giving highest accuracy among other available models
[5]. Such modelled cable system may accurately reflect
behaviour of the protection scheme under various transient
states that are likely to appear. Further detailed information
about establishing computer model of the cable is available in
[5].
Shunt reactors are modelled with series resistance and
inductance parameter values for each phase.
Figure 2. EMTDC/PSCAD Representation of the Underground 420 kV
Cable System.
Table III shows validation results for three possible currents
that can flow through protected cable. Parameters were chosen
that mostly affect proper function of established differential
protection. Detailed methodology for cable model validation
along with series of calculations are given in [1-3].
TABLE III. VALIDATION RESULTS FOR EMTDC/PSCAD UNDERGROUND
CABLE MODEL
Description
Theoretical
Analysis
EMTDC/PSCAD
Model
Relative
error er
(%)
Maximum Charging
Current (kA)
0,878 0,805 8,3
External Fault at ASV
substation (kA)
9,555 11,49 16
Internal Fault in the
middle of the cable (kA)
11,85 12,94 8,4
Relative error is calculated from er = 100· |ITA - IPSCAD| / ITA, where: ITA – current parameter value
obtained algebraically; IPSCAD – current parameter value obtained numerically.
Relative error originates from cable geometry, since mutual
couplings between internal conductive cable layers take place.
This corresponds to core conductors and screens that are in
close proximity to each other. Resulting inductive reactance for
single phase is lower than calculated algebraically, thus giving
higher current value which rises significantly when high
currents flow through cable [2].
Validation results allow to conclude that certain error level
occurs and has to be taken into account. Higher fault current
values from EMTDC/PSCAD simulations allow to keep
safety margin for the analysis based on simulation results.
C. Differential Protection Scheme – Description
Total differential protection scheme for the analyzed cable
system is presented in Fig. 3.
Figure 3. Schematic Representation of the Differential Protection Scheme
over Underground 420 kV Cable System.

cross_A
cross_A
cross_B
PHASE C
A
B
Ctrl
Ctrl = 1
1
EN_1
EN_2
EN_1
EN_2
e-sT
e-sT
e-sT
SAMPLER
PHASE A
PHASE B
cross_B
cross_C
cross_C
trip_C
trip_B
trip_B
trip_C
ib1
ic1
ia1
cross_Ar
cross_Ar
cross_Br A
B
Ctrl
Ctrl = 1
1
B2 cross_Br
cross_Cr
cross_Cr
trip_Cr
trip_Br
trip_Br
trip_Cr
PHASE A
PHASE B
PHASE C
SAMPLER
SAMPLER
SAMPLER
SAMPLER
SAMPLER
ia2
ic2
ib2
trip_Ar trip_Ar trip_A trip_A
e -sT
e -sT
e -sT
B1
SIGNAL
TRANSFORMATION
SIGNAL
TRANSMISSION
SIGNAL
PROCESSING
e
-sT
e
-sT
e
-sT cross_A
cross_A
cross_B
PHASE C
A
B
Ctrl
Ctrl = 1
1
EN_1 EN_1
SAMPLER
PHASE A
PHASE B
cross_B
cross_C
cross_C
trip_C
trip_B
tri p_B
tri p_C
SAMPLER
SAMPLER
trip_A tri p_A
B1
Idiffx
Qdiffxx
Inrush Restraint Ratio
MAXinrush peak
Inrush ON / OFF
1.76
Idiffx
3.52
Idiffxx
0.15
Inrush Restraint Ratio
16.33
MAXinrush peak
1
P1 3.3
P1
*
4.502
Idiffx2 3.52
Idiffx2
Qdiffxx2 7.04
Idiffxx2
*
4.502
Main: Controls
cross-blocking
0
OFF ON


ABC: Graphs
0.040 0.050 0.060 0.070 0.080 0.090 0.100 ...
...
...
0.0
1.0
2.0
y
XX
ABC: Graphs
0.040 0.050 0.060 0.070 0.080 0.090 0.100 ...
...
...
-0.40
-0.30
-0.20
-0.10
0.00
0.10
0.20
0.30
0.40
y
Ia11 Ia11_11 S3 S4 phasor_Ma1 phasor_Pa1 XX
Ia11_1
S / H
in
hol d
out Ia11 Ia11_2
S / H
in
hol d
out Ia11 Ia11_3
S / H
in
hold
out Ia11 Ia11_4
S / H
i n
hold
out Ia11 Ia11_5
S / H
i n
hold
out Ia11 Ia11_6
S / H
i n
hold
out Ia11 Ia11_7
S / H
in
hold
out Ia11 Ia11_8
S / H
in
hol d
out Ia11 Ia11_9
S / H
in
hol d
out Ia11 Ia11_10
S / H
in
hold
out Ia11
Ia11_11
S / H
in
hol d
out Ia11 Ia11_12
S / H
in
hol d
out Ia11 Ia11_13
S / H
in
hold
out Ia11 Ia11_14
S / H
i n
hold
out Ia11 Ia11_15
S / H
i n
hold
out Ia11 Ia11_16
S / H
i n
hold
out Ia11 Ia11_17
S / H
in
hold
out Ia11 Ia11_18
S / H
in
hol d
out Ia11 Ia11_19
S / H
in
hol d
out Ia11 Ia11_20
S / H
in
hold
out Ia11
Ia11_1
S / H
in
hold
out
Ia11

I
a
1
1


I
a
1
1
_
1

D11
D
1
1


Proposed protection scheme consists of following
components:
1) Current Transformers (CT): Devices responsible for
current signal transformation on the level applicable for
measuring instruments installed in protection relays. Table IV
shows current transformers specification used for computer
modelling purpose.
TABLE IV. TECHNICAL DATA OF CURRENT TRANSFORMERS (CT)
Parameter Value
CT manufacturer’s model ABB IMB 420
CT class 5P
Transfomation ratio (A/A) 1000/1
Accuracy Limit Factor – ALF 100
Nominal Power (VA) 15

2) Mono-mode Fibre Optic Cables (FO): Communication
channels responsible for proper signal transmission between
relays. Due to significant length of the protected cable (58,5
km), signal attenuation phenomenon must be considered along
with time delay between sending and reaching signal from
both sides of the protected cable. Reference [2] explains
detailed solution methodology to stated issues. Channel time
delay is calculated based on datasheet provided by fibre optic
cable’s and relay’s manufacturers. Necessary data are gathered
in table V.
TABLE V. TECHNICAL DATA OF FIBRE OPTIC CABLES (FO)
Index Parameter Value
vB Bandwidth Data Speed (bits/s) 512
lFO FO Length (km) 58,5
vFO FO Speed of Light (km/s) 200000
lB HDLC Frame Length (bits) 200

Channel time delay T
delay
is 1,07 ms, calculated from

B
B
FO
FO
B
B
r t s delay
v
l
v
l
v
l
T T T T + + = + + = . (1)
where
T
s
– time for sending signal by the local relay,
T
t
– time for transmitting signal through FO cable,
T
r
– time for receiving signal by the remote relay.
3) Line Differential Relays (DR): Most complex
components realizing signal measurement, signal comparison
and finally - fault detection principles. Relays analyzed in this
paper are Siemens SIPROTEC 4 7SD522. Detailed technical
specification, instruction on establishing proper configuration
parameter set are available in [1], [3] and [4].
D. EMTDC/PSCAD Model of Differential Protection Scheme
A general approach is introduced for protection scheme
modelling in PSCAD software.
Based on previous components description, their unique
characteristic functions are presented on Fig. 4. Each
component is responsible for:
Figure 4. EMTDC/PSCAD Representation of the Differential Protection
Scheme over Underground 420 kV Cable System.
• Signal transformation, modelled by Current
Transformer Lucas model blocks with specified
parameter settings. Reference [6] provides more
information regarding CT Lucas model.
• Signal transmission, modelled by time delay blocks
with specified and calculated time delay value from
(1).
• Signal processing, modelled with complex block
combination, reflecting operation algorithm and
original features of real relays.
III. EMTDC/PSCAD RELAY MODELLING
Relay EMTDC/PSCAD computer model is created in a
shape of box with three phase modules included, so that all
operations are phase segregated as in real relays (see Fig. 7).
Input signals for modules are previously sampled with sampler
blocks, so that 20 sampled values appear each full cycle period
(fixed frequency) [3]. Output logic signal B1 is responsible for
controlling line circuit breaker in case of possible fault
occurrence. Following features are included in each phase
module:
1) Sample Acquisition: Operation necessary for further
phasor and charge computations. Sample values i
n
have to be
stored during full cycle. This operation is available by
implementing 20 Sample/Hold blocks – each controlled by
logic pulse generator block, as presented on Fig. 5. Pulse
generator blocks give command D for each sample/hold block.
Generated pulses are shifted to each other by 18

degrees of
total cycle period.
Figure 5. EMTDC/PSCAD Representation of Sample Acquisition
Technique.

phasor_Ma1 D
+
F
+
Y
X
M X
P Y
M
P
Y
X
M X
P Y
M
P
phasor_Pa1
D
+
F
+ D
+
F
+
X
2
X
2 X
phasor_Ma2
phasor_Pa2 IopA_phasor dyta
Low pass
Butterwth
Order = 1
Sampler
a)













b)

B
+
D
+
F
+
*
*
IresA_phasor
phasor_Ma1
phasor_Ma2
A
B Compar-
ator
A
B Compar-
ator
Idiffx
P1
P1
[Main] Idiffx
[Main] Idiffx2
Edge
Detector
A
B
Ctrl
Ctrl = 1
Idiffx2
Idiffx2
Mono-
T
stable
A
B
Ctrl
Ctrl = 1
EN_1
[Main] EN_1
A
B Compar-
ator
P1
phasor_Ma2 P3
P1 ≥ ≥≥ ≥ phasor_Ma1 P1 < phasor_Ma1

A
B Compar-
ator
P1
phasor_Ma2 P2

2) Phasor Measurement: current phasor values I are
obtained in the shape of complex numbers through Discrete
Fourier Transform technique, based on

C S
I j I I ⋅ + = . (2)
where

⋅ ⋅ ⋅ ⋅ =


=
1
1
∆ ω
2
N
n
n S
i ) t n sin(
N
I . (3)

⋅ ⋅ ⋅ + + ⋅ =


=
1
1
0
∆ ω
2 2
2
N
n
n
N
C
i ) t n cos(
i i
N
I . (4)
Herein are
n = 1,2,…,20 – sample number,
i
n
– current sample value corresponding to sample n,
ω = 2⋅f⋅π – cycle pulsation,
∆t = (f ⋅ N)
-1
– sample time interval,
f = 50 Hz – frequency,
N = 20 – number of samples over one cycle.
Equations (3) and (4) are realized by correlating sample
values with sine and cosine waveforms and summating them
each full cycle period [1].
3) Charge Measurement: Charge values Q are obtained
based on

i
n
n
n
n i
i
t i dt i Q ∆
5
0
5
⋅ ≈ ⋅ =


+
+
+
=
. (5)
Four charge values are calculated each full cycle period. By
applying signal switch block, final charge signal is switched
each quarter cycle. This corresponds to real relay feature,
where charge comparison is performed four times more often
than phasor comparison.
4) Phasor Comparison: based on relay’s principles given
in [1], values for operational phasor I
OP
and restraint phasor
I
RES
are obtained and relay operating criterion is

RES OP
I I > , (6)
for
¹
´
¦
> ⋅

+
¹
´
¦
> ⋅

+ =
>
1 3
1 ≤ 2
1 3
1 ≤ 2
2 2
2 2
1 1
1 1
P I : I P
P I : I P
P I : I P
P I : I P
I I
diff RES
. (7)

2 1
I I I
OP
+ = . (8)
Relay setting parameters P1, P2, P3 and I
diff>
are chosen
based on procedure given in [3] and [4]. Parameters I
1
and I
2

are current phasor values correspondingly measured by local
and remote relays. Operation of switching multiplying factors
for restraint phasor (determined by current signal value – in
fault state or load state) is made with the usage of comparator
blocks, which output signal is multiplied by its corresponding
actual phasor current signal value, as shown on Fig. 6.
Based on information obtained from the position of line
circuit breaker installed on the same side as device, differential
relay can detect “dead line” state when no current flows
through the protected cable. Cable energization state – when
circuit breaker is suddenly switched on – is detected by Edge
Detector block by positive transition appearance of signal from
line circuit breakers. This allows generating digital impulse,
which is later extended to the specified time interval - Td
setting, which can be changed based on relay settings by
Monostable Multivibrator block.
Figure 6. EMTDC/PSCAD Representation of Phasor Comparison technique:
a) Restraint phasor IRES, b) Operational phasor IOP.
5) Charge Comparison: for this technique, the same
algorithm is used as for phasor comparison. I
diff>
parameter is
replaced with minimum threshold value for charges: I
diff>>
. In
addition, phasor signals are replaced with their corresponding
operational Q
OP
and restraint Q
RES
charge values.
6) Signal Filtering: operational and restraint values are
filtered using low-pass Butterworth filter block with
established frequency threshold corresponding to each
comparison technique.
7) Inrush Restraint: 2nd harmonic phasor currents I
2nd
are
measured by online frequency scanner blocks. If its values
exceed established ratio k
ratio
of 1st harmonic I
1st
, relay
prevents tripping operation. In EMTDC/PSCAD model this
feature can be switched OFF as in real relays. Condition
statement (9) has to be fulfilled in order to activate inrush
restraint blocking feature. Upper limit for non-tripping
operation is established with I
max_peak
parameter
( ) ( )
peak max_
st st
ratio
nd
I I I k I < ∧ ⋅ >
1 1 2
. (9)
Comparison principles are obtained with a combination of
comparator blocks. Output signals from comparators can then

cross_A
cross_A
cross_B
PHASE C
A
B
Ctrl
Ctrl = 1
1
EN_1
EN_1
SAMPLER
PHASE A
PHASE B
cross_B
cross_C
cross_C
trip_C
trip_B
trip_B
trip_C
SAMPLER
SAMPLER
trip_A trip_A
B1

PHASE_C: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.0
3.0
6.0
9.0
12.0
15.0
A
IopC_phasor IresC_phasor IopC_charge IresC_charge
PHASE_B_1 : Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.0
3.0
6.0
9.0
12.0
15.0
A
IopB_phasor IresB_phasor IopB_charge IresB_charge
PHASE_A_1 : Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.0
3.0
6.0
9.0
12.0
15.0
A
IopA_phasor IresA_phasor QopA_charge QresA_charge
PHASE_C: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240
0.0
6.0
12.0
18.0
24.0
30.0
36.0
42.0
48.0
54.0
60.0
66.0
A
IopC_phasor IresC_phasor
PHASE_B_1: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240
0.0
6.0
12.0
18.0
24.0
30.0
36.0
42.0
48.0
54.0
60.0
66.0
A
IopB_phasor IresB_phasor
PHASE_A_1 : Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240
0.0
6.0
12.0
18.0
24.0
30.0
36.0
42.0
48.0
54.0
60.0
66.0
A
IopA_phasor IresA_phasor


r
e
s
t
r
a
i
n
t

p
h
a
s
o
r

I
R
E
S






o
p
e
r
a
t
i
o
n
a
l

p
h
a
s
o
r

I
O
P




C

p
h
a
s
e


(
A
)












B

p
h
a
s
e


(
A
)











A

p
h
a
s
e

(
A
)




r
e
s
t
r
a
i
n
t

c
h
a
r
g
e

Q
R
E
S






o
p
e
r
a
t
i
o
n
a
l

c
h
a
r
g
e

Q
O
P






C

p
h
a
s
e


(
A
·
s
)








B

p
h
a
s
e


(
A
·
s
)








A

p
h
a
s
e

(
A
·
s
)







a) b)
Time (s) Time (s)

IRES
IOP
QRES
QOP


PHASE_C: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
A
IopC_phasor IresC_phasor IopC_charge IresC_charge
PHASE_B_1: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
A
IopB_phasor IresB_phasor IopB_charge IresB_charge
PHASE_A_1 : Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280 0.320 0.360 0.400 0.440 0.480 ...
...
...
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
A
IopA_phasor IresA_phasor QopA_charge QresA_charge
PHASE_C: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
A
IopC_phasor IresC_phasor IopC_charge
PHASE_B_1 : Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
A
IopB_phasor IresB_phasor IopB_charge
PHASE_A_1: Graphs
0.000 0.040 0.080 0.120 0.160 0.200 0.240 0.280
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
A
IopA_phasor IresA_phasor QopA_charge


r
e
s
t
r
a
i
n
t

p
h
a
s
o
r

I
R
E
S






o
p
e
r
a
t
i
o
n
a
l

p
h
a
s
o
r

I
O
P




C

p
h
a
s
e


(
A
)












B

p
h
a
s
e


(
A
)











A

p
h
a
s
e

(
A
)




r
e
s
t
r
a
i
n
t

c
h
a
r
g
e

Q
R
E
S






o
p
e
r
a
t
i
o
n
a
l

c
h
a
r
g
e

Q
O
P






C

p
h
a
s
e


(
A
·
s
)








B

p
h
a
s
e


(
A
·
s
)








A

p
h
a
s
e

(
A
·
s
)







a) b)
Time (s) Time (s)
IRES
IOP QRES
QOP
be combined with logic gates so that tripping signal depends on
the resulting signal from the inrush restraint feature.
8) Cross-blocking: in order to prevent tripping signals
from all three phases when inrush feature is active in only
single phase, cross-blocking feature is introduced. Its PSCAD
representation is shown on Fig. 7.
Figure 7. EMTDC/PSCAD Representation of Cross-Blocking Technique.
In EMTDC/PSCAD computer model, cross-blocking
utilizes single phase tripping and inrush activation signals as
the output signals of each phase module. Combining them all
with logic gates gives final tripping signal decision B1. Hence,
described feature has to be implemented outside phase
modules. As in real relays, feature can be permanently
switched OFF during normal operation.
Original EMTDC/PSCAD files with fully established and
configured models of relays and protected cable system are
available at main author on request.
IV. EMTDC/PSCAD SIMULATION CASES
A. Two-Phase External Fault at KYV substation
External fault simulation in phases A and B allows analysis
on how relay computer model reacts when high currents flow
through the protected cable. Fault is cleared after 55 ms by
virtual bus protection installed in place where fault occurred.
All shunt reactors are disconnected (highest charging current).
Simulation graphs are presented on Fig. 8.
Figure 8. External Fault State at 10 ms: a) Phasor Comparison Technique, b)
Charge Comparison Technique.
Due to high currents flowing through phases A and B
which are higher than calculated P1 value [4], transition takes
place resulting in switching multiplying factors from P2 to P3
value. This means that transformed secondary current lies
within fault area and security margin is increased in
corresponding phases. On presented plots, restraint threshold is
higher for the time when external fault current flows. After
fault clearing, restraint values return to its normal threshold
levels since transformed current lies once again within load
area. An increase of operational values in phases with high
current appears after fault clearing, giving large safety margin
in order to prevent unwanted tripping. It is seen that during
whole simulation operational values do not exceed restraint
ones. As a result, relay properly does not detect any fault
within protected cable and does not send tripping signal.
B. Single-Phase Intenal Fault at KYV substation
Simulation test involves internal fault appearance in phase
A within protected cable. As the worst case scenario, single A-
phase fault is applied with high resistance R
fault
= 20 Ω and all
shunt reactors are switched ON (lowest charging current).
Computer model with established setting parameters should be
able to properly detect and recognize fault state within phases.
Figure 9 presents described simulation case results .
Figure 9. Internal Fault State at 10 ms: a) Phasor Comparison Technique, b)
Charge Comparison Technique.
As expected, internal fault occurred in A phase and is
detected by relay computer model both with phasor and charge
comparison techniques. Operational values significantly exceed
restraint ones after 20 ms from fault occurrence for phasor and
15ms for charge comparison principles. Earlier fault detection
with charge technique results in sending tripping signal after
15ms in order to disconnect faulted cable.
V. RELAY TESTING
A. Description
Relay experimental testing is possible with the usage of
modern equipment and software capable of converting current
signals from EMTDC/PSCAD software into current
waveforms injected into real differential relays. Simplified
diagram of experimental test setup is presented on Fig. 10.
U
DC
supply
Ia11
Ib11
Ic11
FO cable
DIGSI 4.81
Advanced Transplay
UDC supply
Ia22
Ib22
Ic22


Figure 10. Experimental Test Setup.
Six current signals are sent: Ia11, Ib11, Ic11, Ia22, Ib22,
Ic22 from which three enter to each relay accordingly to the
side from which they were measured. Relays interconnected
together with fibre optic cable, respond based on delivered
signals with measured values and annunciation messages saved
as logs. These logs can then be sent to PC and read in DIGSI
software for further analysis and for comparison purposes.
B. Results
All tests from experimental analysis and EMTDC/PSCAD
simulations were performed with the same setting parameter
values. Relay’s operating speed and sensitivity have been
examined.
1) Operating Speed: Operating speed analysis gives idea
on how fast relay is able to detect fault states. By the analysis
of restraint/operational plots in EMTDC/PSCAD computer
software, time interval between exceeding threshold by
operational charge value Q
OP
and phasor value I
OP
can be
compared with the ones obtained from DIGSI logs. Analyzed
study case results are presented on table VI.
TABLE VI. VALIDATION RESULTS OF OPERATING SPEED FOR
EMTDC/PSCAD RELAY MODEL
Time interval (ms)
Case description
Experimental
Results
PSCAD
Simulation Results
Single-phase to ground internal
fault in the middle
8 14
Single-phase to ground internal
fault at KYV busbar
8 14
Two-phase to ground internal
fault at KYV busbar
14 │ 14 15 │ 15
Three-phase to ground internal
fault KYV busbar
18 │ 12 │ 18 15 │ 10 │ 15

2) Sensitivity: Relay’s sensitivity analysis is critical for
proper internal fault states recognition. For this reason,
internal faults with very high resistance values were analyzed.
Differential threshold parameter I
diff>
for phasor comparison
was adjusted in order to obtain its critical threshold values.
Results are listed and compared in table VII. Phasor
comparison is examined since it is more sensitive and
necessary for proper fault detection. Critical values are these
on which relay still detects fault and - if increased of a single
setting step - makes no reaction for the same fault conditions.

TABLE VII. VALIDATION RESULTS OF SENSITIVITY FOR EMTDC/PSCAD
RELAY MODEL
Differential phasor Idiff>
threshold for tripping
Differential phasor Idiff>
threshold for non-tripping
Internal
fault
resistance
(Ω)
Experimental
Results
PSCAD
Simulation
Results
Experimental
Results
PSCAD
Simulation
Results
55 4,02 4,05 4,03 4,06
70 3,23 3,25 3,24 3,26
145 1,72 1,73 1,73 1,74
210 1,31 1,32 1,32 1,33

VI. CONCLUSIONS
EMTDC/PSCAD relay computer model proves to be
reliable and efficient from taken simulation cases with
established parameter set. Apart from internal and external
fault states, analyzed simulation cases included transmission
cable’s energization and shunt reactor’s energization states,
giving overall 10 different study cases [2]. All simulation
results have been successfully compliant with the expected
ones.
According to simulation results, relay model is able to
accurately detect internal faults and differentiate them with
mentioned other states that may be misleading. Very high fault
resistances from which relay cannot detect faults marks
efficiency of specific algorithms implemented and used for
measurement and comparison purposes of the obtained signals.
VII. ACKNOWLEDGMENT
The first author gratefully acknowledges research support
from the Danish TSO - Energinet.dk, which delivered all
necessary technical data of the analyzed system along with
Siemens SIPROTEC 4 7SD522 relays.
VIII. REFERENCES
[1] G. Ziegler, Numerical Differential Protection: Principles and
Applications, Nuremberg, Germany: Siemens AG, Jul. 2005.
[2] M. Sztykiel, “Protection philosophies for HVAC transmission network,”
M.S. thesis, IET, Aalborg Univ., Aalborg, Denmark, 2009.
[3] SIPROTEC 4 Differential Protection 7SD52 V4.1 Manual, Siemens AG,
Nuremberg, Germany, 2002.
[4] T. Sezi at al., “Field Experience Summary with a Line Differential
Relay Using Complex Communication Infrastructure,” Siemens Energy,
Inc., Nuremberg, Germany, Sep. 2009.
[5] Applications of PSCAD/EMTDC Application Guide, Manitoba HVDC
Research Centre Inc., Winnipeg, Canada.
[6] J. Rohan Lucas, “Representation of Magnetisation Curves over a wide
region using a non-integer power series,” IJEEE: Manchester Univ.
Press, vol. 25, No 4, pp. 335-340, Oct. 1988.

78 [ohm] Ip1 C A B Phase Currents Waveforms .001 [ohm] 0.0 185 127.61 [H]2. Detailed information about cable structure is presented on table I.159 [ohm] A B Ph C 4.001 [ohm] 0..001 [ohm] XLPE8 C1 S1 C2 S2 C3 S3 C1XLPE7 S1 C2 S2 C3 S3 0.858 [ohm] A B A C B 4. 2.00 -0. Figure 3.001 [ohm] 0.61 [H]2.0 6.175 0. .0 [Hz] 10477 [MVA] Z1 = 0. four switchable shunt reactors are installed.6045 Q_ASV 6 5 4 3 2 1 Phase Angle_2 180 TOR1r ASVr -180 kA degrees 0 kV 0 0. and earthed each 6 km.159 [ohm] C XLPE6 TOR1 shunt reactor C XLPE10 C XLPE9 C XLPE8 C XLPE7 C1XLPE10 A Ip2 EN_2 B C C3 ASVr C2 S2 S1 0.555 11. 3) Supply Sources (ES1–ES2): Power system on both sides of the cable is modelled by ES1 and ES2 sources that are Thevenin equivalents consisting of voltage sources and its short-circuit impedances.0 [kV].54 444.0 -6.80 10.60 0.001 [ohm] P_ASV 5.001 [ohm] XLPE C1 S1 C2 S2 C3 S3 C1XLPE2 S1 C2 S2 C3 S3 0.40 0.829 [ohm] + j16.0 8.001 [ohm] 0. Resulting inductive reactance for single phase is lower than calculated algebraically. Parameters were chosen that mostly affect proper function of established differential protection. C.80 0.The system consists of the following components: 1) XLPE HVAC Underground Cable Sections (L1–L2): Selected cable is made of three aluminium single-core cables buried underground on the depth of 1.001 [ohm] C3 S3 0. TECHNICAL DATA OF XLPE UNDERGROUND 420 KV CABLE Description Cross-section of conductor (mm2) Diameter of conductor (mm) Insulation thickness (mm) Diameter over insulation (mm) Cross-section of screen (mm2) Outer diameter of cable (mm) Capacitance (µF/km) Inductance (mH/km) Charging current per phase (A/km) Value 1600 52 27.001 [ohm ] TOR1r 5.001 [ohm] 0.829 + j16.001 [ohm ] XLPE4 C1 S1 C2 S2 C3 S3 C1XLPE5 S1 C2 S2 C3 S3 0.0 -2.050 0.001 [ohm] 0.001 [ohm] 0.025 0.21 0.IPSCAD| / ITA.001 [ohm] 0.001 [ohm] XLPE7 C1 S1 C2 S2 C3 S3 C1XLPE6 S1 C2 S2 C3 S3 0. DR ASV Protected Zone CT CT FO DR KYV Supply Source ES1 ES2 B.001 [ohm] XLPE3 C1 S1 C2 S2 C3 S3 C1XLPE4 S1 C2 S2 C3 S3 0.0 -10. Detailed methodology for cable model validation along with series of calculations are given in [1-3].001 [ohm] XLPE5 C1 S1 C2 S2 C1XLPE6 A B C A B C C3 S3 C2 S2 S1 0.61 [H]2. C XLPE C XLPE2 C XLPE3 C XLPE4 C XLPE5 C1XLPE A Ip1 EN_1 C B C2 S2 C3 KYVr S1 0.60 -0.0 [Hz] 10592 [MVA] Z1 = 0.0 110.01 [H]2.01 [H]2. Higher fault current values from EMTDC/PSCAD simulations allow to keep safety margin for the analysis based on simulation results.001 [ohm] 0.80 -1.61 [H]2.50 14. 50.839 + j16.001 [ohm] XLPE2 C1 S1 C2 S2 C3 S3 C1XLPE3 S1 C2 S2 C3 S3 0.200 0 0 400 400 1304.80 0. TABLE I.40 0. 50.0 4. Table III shows validation results for three possible currents that can flow through protected cable.01 [H]2. each on ASV.159 [ohm] 4.159 [ohm] 4. Metallic screens of each cable section are crossbonded approximately each 2 km. thus giving higher current value which rises significantly when high currents flow through cable [2].20 0.001 [ohm] 0.20 0. described cable system is modelled in EMTDC/PSCAD software by frequency dependent (phase) model.878 9. 3.858 [ohm] 400.3 16 8.001 [ohm] 0.94 Relative error er (%) 8.0 2.0 [kV].001 [ohm ] KYV source P_KYV KYV shunt reactor S3 0.00 0.5 km.3 m and laid in a flat formation within 300 mm from each other.01 [H]2.159 [ohm] 4. where: ITA – current parameter value obtained algebraically.0 -4. Further detailed information about establishing computer model of the cable is available in [5].001 [ohm] ASV source ASV shunt reactor S3 0. Figure 2..01 [H]2. giving highest accuracy among other available models [5].805 11. since mutual couplings between internal conductive cable layers take place.61 [H]2.6 [ohm] 5. Differential Protection Scheme – Description Total differential protection scheme for the analyzed cable system is presented in Fig.159 [ohm] V C Q_KYV TOR2r TOR2 shunt reactor Ip2 Ph C A B A A Logic Fault Tim ed FAULTS B C EXTERNA FA L ULT INTERNAL FAULT 0..0 0.60 0.9 Shunt reactors are modelled with series resistance and inductance parameter values for each phase. Such modelled cable system may accurately reflect behaviour of the protection scheme under various transient states that are likely to appear.858 [ohm] Q_ASV 400. Validation results allow to conclude that certain error level occurs and has to be taken into account.both line ends M : Phase currents ain 4.001 [ohm] 0.498 1 .61 [H]2.4 Description Maximum Charging Current (kA) External Fault at ASV substation (kA) Internal Fault in the middle of the cable (kA) 2) Shunt Reactor Banks (SR1–SR4): For reactive power compensation..40 -0. TABLE III. KYV bus bars with reactive power of 100 MVARs and two between cable sections with reactive power of 140 MVARs.100 0.78 Relative error is calculated from er = 100· |ITA . Total cable is divided into two sections of lengths accordingly 28 km and 29. TABLE II.001 [ohm] 0.0 -8..20 -0.. EMTDC/PSCAD Model of Cable System In Fig.49 12.858 [ohm] B 5.0 0.000 Ia1 Ia2 V kA Ia1 Ia2 Sources Control Set Timed Fault Logic Faults Control Set Faults Control Set Description Fault Type Control ------------------------------1 = No Fault 2 = Phase Ato Ground 3 = Phase B to Ground 4 = Phase C to Ground 5 = Phase A to Ground B 6 = Phase A to Ground C 7 = Phase BC to Ground 8 = Phase A to Ground BC 9 = Phase A B 10 = Phase ABC kA KYV Phase Angle_1 180 Faults : C ontrols Voltage_1 P_KYV Q_KYV External Fault 10 9 8 7 Faults : C ontrols Internal Fault 10 9 8 7 6 5 4 3 2 1 Phase Angle_1 Voltage_1 500 degrees Line Circuit Breakers Control Set Shunt Reactors Circuit Breakers Control Set Timed Timed Breaker Breaker Logic Logic EN_1 Clos ed@t0 Closed@t0 Timed Timed Breaker Breaker Logic Logic KYVr Open@t0 Open@t0 Timed Breaker Logic Open@t0 Timed Breaker Logic Open@t0 Ia1 Ia2 kV EN_2 -180 0 0 TOR2r 0 Phase Angle_2 400 ASV Voltage_2 500 400 Voltage_2 944 P_ASV 82. IPSCAD – current parameter value obtained numerically.125 0.20 -0.001 [ohm ] 0.001 [ohm] 0. . This corresponds to core conductors and screens that are in close proximity to each other.001 [ohm ] XLPE9 C1 S1 C2 S2 C3 S3 C1XLPE8 S1 C2 S2 C3 S3 0.001 [ohm] 0.85 EMTDC/PSCAD Model 0.858 [ohm] 5. . Relative error originates from cable geometry.858 [ohm] C 5.40 -0.001 [ohm] 0. Schematic Representation of the Differential Protection Scheme over Underground 420 kV Cable System.01 [H]2.001 [ohm ] 0.001 [ohm ] 0.001 [ohm] XLPE10 C1 S1 C2 S2 C3 S3 C1XLPE9 S1 C2 S2 C3 S3 0.001 [ohm ] 0.60 -0. Parameter values are listed on table II.150 0. TECHNICAL DATA OF CABLE’S SUPPLY SOURCES Voltage (kV) 420 420 Short-circuit impedance ( ) 0. VALIDATION RESULTS FOR EMTDC/PSCAD UNDERGROUND CABLE MODEL Theoretical Analysis 0.60 0.075 0.0 0.001 [ohm] 0. EMTDC/PSCAD Representation of the Underground 420 kV Cable System.00 -0.001 [ohm ] 0.839 [ohm] + j16.

3) Line Differential Relays (DR): Most complex components realizing signal measurement.080 0.100 .10 0.00 -0. as presented on Fig.502 Qdiffxx Qdiffxx2 Inrus h Res traint Ratio MAX inrush peak Inrus h ON / OFF trip_A cros s_A trip_B cross _A cros s_B trip_C cros s_C cross _B cross _C 1 B Ctrl EN_2 EN_1 E N_2 E N_1 Ctrl A Ctrl = 1 B1 B2 1 = Ctrl A B 1 cros s_Ar cros s_Br cros s_Cr trip_Cr cros s_Cr cros s_Br trip_A trip_B trip_C M : Controls ain cross-blocking OFF ON -s T e -s T e -s T e 0 SAMPLER ia1 PHAS E A trip_Ar trip_Br trip_Cr trip_Ar cros s_Ar trip_Br PHASE A S AMPLER ia2 PHAS E B SAMPLER ib1 PHASE B S AMPLER ib2 PHASE C SAMPLER ic1 PHASE C S AMPLER ic2 0.050 0.050 0.20 0. Output logic signal B1 is responsible for controlling line circuit breaker in case of possible fault occurrence. EMTDC/PSCAD Model of Differential Protection Scheme A general approach is introduced for protection scheme modelling in PSCAD software. TABLE IV. EMTDC/PSCAD Representation of Sample Acquisition Technique.10 -0. 5.07 ms. modelled by time delay blocks with specified and calculated time delay value from (1).100 . III.. . Relays analyzed in this paper are Siemens SIPROTEC 4 7SD522.fault detection principles.090 0... modelled with complex block combination. signal comparison and finally . This operation is available by implementing 20 Sample/Hold blocks – each controlled by logic pulse generator block.3 P1 1.. Signal transmission.15 Inrus h Restraint Ratio 16. so that 20 sampled values appear each full cycle period (fixed frequency) [3]. Reference [2] explains detailed solution methodology to stated issues. their unique characteristic functions are presented on Fig. ..52 Idiffx2 3.20 -0. signal attenuation phenomenon must be considered along with time delay between sending and reaching signal from both sides of the protected cable. .. Ia11_1 D11 D11 y y 1. vB vFO vB Relay EMTDC/PSCAD computer model is created in a shape of box with three phase modules included. Channel time delay is calculated based on datasheet provided by fibre optic cable’s and relay’s manufacturers. TECHNICAL DATA OF CURRENT TRANSFORMERS (CT) Parameter CT manufacturer’s model CT class Transfomation ratio (A/A) Accuracy Limit Factor – ALF Nominal Power (VA) Value ABB IMB 420 5P 1000/1 100 15 Based on previous components description. TABLE V. • 2) Mono-mode Fibre Optic Cables (FO): Communication channels responsible for proper signal transmission between relays. EMTDC/PSCAD Representation of the Differential Protection Scheme over Underground 420 kV Cable System. so that all operations are phase segregated as in real relays (see Fig.30 (1) where Ts – time for sending signal by the local relay. D. Generated pulses are shifted to each other by 18 degrees of total cycle period.502 * 4. 7).0 0. instruction on establishing proper configuration parameter set are available in [1]..04 Idiffxx2 P1 Idiffx Idiffx2 * 4.040 XX 0..060 0.080 0. 4.Proposed components: protection scheme consists of following 1) Current Transformers (CT): Devices responsible for current signal transformation on the level applicable for measuring instruments installed in protection relays. Each component is responsible for: 3. Pulse generator blocks give command D for each sample/hold block.40 Ia11 Ia11_11 S3 S4 phasor_M a1 phasor_P a1 XX 0.0 0. Due to significant length of the protected cable (58. Ia11 S/H in out hold ●Ia11 ●Ia11_1 0.0 0.5 km). reflecting operation algorithm and original features of real relays.30 -0. EMTDC/PSCAD RELAY MODELLING • • Channel time delay Tdelay is 1. Table IV shows current transformers specification used for computer modelling purpose. Input signals for modules are previously sampled with sampler blocks. Necessary data are gathered in table V. Sample values in have to be stored during full cycle.060 0. Ia11 S/H in out hold Ia11_1 Ia11 S/H in out hold Ia11_2 Ia11 S/H in out hold Ia11_3 Ia11 S/H in out hold Ia11_4 Ia11 S/H in out hold Ia11_5 Ia11 S /H in out hold Ia11_6 Ia11 S /H in out hold Ia11_7 Ia11 S/H in out hold Ia11_8 Ia11 S/H in out hold Ia11_9 Ia11 S/H in out hold Ia11_10 Ia11 S/H in out Ia11_11 Ia11 hold S/H in out Ia11_12 Ia11 hold S/H in out Ia11_13 Ia11 hold S/H in out Ia11_14 Ia11 hold S/H in out Ia11_15 Ia11 hold S /H in out Ia11_16 Ia11 hold S /H in out Ia11_17 Ia11 hold S/H in out Ia11_18 Ia11 hold S/H in out Ia11_19 Ia11 hold S/H in out Ia11_20 hold Figure 5.52 Idiffxx 7..5 200000 200 Signal transformation. .090 0.070 ABC : G raphs 2.070 0. .. Detailed technical specification.040 0. Following features are included in each phase module: 1) Sample Acquisition: Operation necessary for further phasor and charge computations.76 Idiffx 3. calculated from Tdelay = Ts + Tt + Tr = lB lFO lB + + . ABC : G raphs 0.33 MAX inrus h peak 1 e-s T e-s T e-s T EN_1 EN_1 PHASE A tri p_A cross _A tri p_B trip_A trip_B trip_C -sT e -sT e -sT e SAMPLER PHASE B SAMPLER cross_A cross _B cross_B cross_C 1 A B Ctrl = 1 B1 PHASE C SAMPLER tri p_C cross _C Ctrl SIGNAL TRANSFORMATION SIGNAL TRANSMISSION SIGNAL PROCESSING Figure 4. Tt – time for transmitting signal through FO cable. [3] and [4].. Signal processing. modelled by Current Transformer Lucas model blocks with specified parameter settings. Tr – time for receiving signal by the remote relay. Reference [6] provides more information regarding CT Lucas model.40 0. Index vB lFO vFO lB TECHNICAL DATA OF FIBRE OPTIC CABLES (FO) Parameter Bandwidth Data Speed (bits/s) FO Length (km) FO Speed of Light (km/s) HDLC Frame Length (bits) Value 512 58..

Output signals from comparators can then . (7) ( I 2nd > k ratio ⋅ I 1st ) ∧ ( I 1st < I max_ peak ) . By applying signal switch block. which can be changed based on relay settings by Monostable Multivibrator block. final charge signal is switched each quarter cycle. Comparison principles are obtained with a combination of comparator blocks. If its values exceed established ratio kratio of 1st harmonic I1st.20 – sample number. N = 20 – number of samples over one cycle. f = 50 Hz – frequency. In addition. phasor signals are replaced with their corresponding operational QOP and restraint QRES charge values. Based on information obtained from the position of line circuit breaker installed on the same side as device. This allows generating digital impulse. EMTDC/PSCAD Representation of Phasor Comparison technique: a) Restraint phasor IRES. where charge comparison is performed four times more often than phasor comparison. b) Operational phasor IOP. In EMTDC/PSCAD model this feature can be switched OFF as in real relays.2) Phasor Measurement: current phasor values I are obtained in the shape of complex numbers through Discrete Fourier Transform technique. a) P1 ≥ phasor_Ma1 A phas or_Ma2 P1 B Comparator P2 phas or_Ma2 P1 B Comparator (3) IC = Herein are 2 N N −1 i i  ⋅  0 + N + cos( ω ⋅ n ⋅ ∆t ) ⋅ in  . values for operational phasor IOP and restraint phasor IRES are obtained and relay operating criterion is I OP > I RES . P2. Equations (3) and (4) are realized by correlating sample values with sine and cosine waveforms and summating them each full cycle period [1]. 4) Phasor Comparison: based on relay’s principles given in [1].…. differential relay can detect “dead line” state when no current flows through the protected cable. 6) Signal Filtering: operational and restraint values are filtered using low-pass Butterworth filter block with established frequency threshold corresponding to each comparison technique. in – current sample value corresponding to sample n.Td setting. (8) (9) I OP = I 1 + I 2 . 3) Charge Measurement: Charge values Q are obtained based on 5+ n P1 < phasor_Ma1 A P3 A phasor_Ma1 Idiffx2 A B Ctrl B Idiffx2 Ctrl P1 Ctrl = 1 B Comparator * [Main] Idiffx2 [Main] Idiffx A Idiffx Ctrl = 1 [Main] EN_1 EN_1 Edge Detector Monostable D T B + + + F A * IresA_phasor phasor_Ma2 P1 B Comparator b) phas or_Ma1 phas or_Pa1 phas or_Ma2 phas or_Pa2 M Y P P M Y P P M X X + D + Y X F 2 X M X Y D + F + dyta 2 X D + F + X Low pas s Butterwth Order = 1 Sampler IopA_phas or Q= 0+ n ∫ i ⋅ dt ≈ ∑ i ⋅ ∆t . 6. ω = 2⋅f⋅π – cycle pulsation. as shown on Fig. based on I = I S + j ⋅ IC . 2 2 n =1   ∑ (4) n = 1. the same algorithm is used as for phasor comparison. where IS = 2 N  N −1  ⋅ ∑ sin( ω ⋅ n ⋅ ∆t ) ⋅ in  . Idiff> parameter is replaced with minimum threshold value for charges: Idiff>>. This corresponds to real relay feature. which output signal is multiplied by its corresponding actual phasor current signal value. 7) Inrush Restraint: 2nd harmonic phasor currents I2nd are measured by online frequency scanner blocks. Upper limit for non-tripping operation is established with Imax_peak parameter (6) for P2 ⋅ I 1 : I RES = I diff > +   P3 ⋅ I 1 : I 1 ≤ P1  P 2 ⋅ I 2 : + I 1 > P1  P3 ⋅ I 2 : I 2 ≤ P1 I 2 > P1 . Cable energization state – when circuit breaker is suddenly switched on – is detected by Edge Detector block by positive transition appearance of signal from line circuit breakers. (5) Four charge values are calculated each full cycle period. ∆t = (f ⋅ N)-1 – sample time interval. Condition statement (9) has to be fulfilled in order to activate inrush restraint blocking feature. i i i=n n +5 Figure 6.  n =1  (2) Relay setting parameters P1. P3 and Idiff> are chosen based on procedure given in [3] and [4].2. Parameters I1 and I2 are current phasor values correspondingly measured by local and remote relays. Operation of switching multiplying factors for restraint phasor (determined by current signal value – in fault state or load state) is made with the usage of comparator blocks. relay prevents tripping operation. which is later extended to the specified time interval . 5) Charge Comparison: for this technique.

0 3.480 0.440 0..00 4.240 0.080 . 8) Cross-blocking: in order to prevent tripping signals from all three phases when inrush feature is active in only single phase.080 .000 0.00 3..0 36..080 . Combining them all with logic gates gives final tripping signal decision B1.00 3.50 2.0 35.240 0. As in real relays.120 0. feature can be permanently switched OFF during normal operation.00 0.280 0.200 0..0 18.440 0.50 2. V. 0.200 0.0 IresB_phasor A.080 0.400 0.0 12. 0.50 1. Simulation graphs are presented on Fig. b) Charge Comparison Technique. Its PSCAD representation is shown on Fig..320 IRES IOP QRES QOP 0.0 A 20.00 2.160 0.0 3. Earlier fault detection with charge technique results in sending tripping signal after 15ms in order to disconnect faulted cable.0 42.0 60.120 0.00 3.080 0.120 0.240 0.400 0. .0 0.0 0. Figure 7.50 4. 0. After fault clearing. An increase of operational values in phases with high current appears after fault clearing. EMTDC/PSCAD SIMULATION CASES PH ASE _B_1 : Graphs 15.280 0.00 4.200 0.0 25..160 0.0 0.200 0.440 B phase (A) 5.0 9.480 .280 0.00 2.0 0. All shunt reactors are disconnected (highest charging current).0 60.00 0. restraint threshold is higher for the time when external fault current flows...040 0.160 0.000 0.00 1.480 0.000 0.0 0.080 0.0 A 6.360 Time (s) Time (s) Figure 8. internal fault occurred in A phase and is detected by relay computer model both with phasor and charge comparison techniques.000 IopB_phasor IresB_phasor P ASE H _B_1 : Graphs IopB_charge IresB_charge 40.000 0.120 0. Internal Fault State at 10 ms: a) Phasor Comparison Technique.00 0.480 0.0 3. 0.0 0.120 0.0 25.0 12. 8. relay properly does not detect any fault within protected cable and does not send tripping signal.240 0.0 60.200 0.00 2.000 0. Fault is cleared after 55 ms by virtual bus protection installed in place where fault occurred.080 0.360 0.240 Time (s) ● restraint charge QRES │ ● operational charge QOP IopA_phasor IresA_phasor QopA_charge A phase (A·s) A phase (A) a) P HASE _A_1 : Graphs b) P HASE _A_1 : G raphs QresA_charge 40.280 Figure 9.50 3.000 0..160 0.360 0.080 .040 .160 0. described feature has to be implemented outside phase modules. b) Charge Comparison Technique.000 0.320 0.280 0.. Figure 9 presents described simulation case results .280 IopC_phasor IresC_phasor IopC _charge As expected. 0.0 0. .0 36.0 42.040 0.50 0.0 IopC_phasor IresC_phasor A 0..040 0.400 0.280 C phase (A) 5. EN_1 EN_1 PHASE A trip_A cross_A trip_B cross_B trip_A trip_B trip_C SAMPLER PHASE B SAMPLER cross_A cross_B cross_C 1 A B Ctrl = 1 B1 PHASE C SAMPLER trip_C cross_C Ctrl value.0 IopC_phasor IresC_phasor IopC_charge P HASE : G _C raphs C phase (A·s) IresC_charge C phase (A) 12.00 0..440 0.0 18.0 35.400 0.0 30.0 24.50 0...320 0.0 25.160 0...360 0.200 0.0 42.400 0.160 0. EMTDC/PSCAD Representation of Cross-Blocking Technique.0 6.080 . IopB_phasor 0.040 Time (s) .50 3.040 0. .50 0.040 0.0 0.160 0. Description Relay experimental testing is possible with the usage of modern equipment and software capable of converting current signals from EMTDC/PSCAD software into current waveforms injected into real differential relays..00 1. On presented plots.0 A 20.00 4.0 IopA_phasor IresA_phasor QopA_charge QOP QRES . .0 IopB_phasor IresB_phasor P ASE H _B_1 : G raphs IopB_charge 0..280 0. Due to high currents flowing through phases A and B which are higher than calculated P1 value [4]. giving large safety margin in order to prevent unwanted tripping.280 0.0 15.0 5.000 0.200 0. 10.50 3..0 54.440 0.200 0.50 1.0 24.0 10.00 0.00 A 0.0 6.0 24.0 30.360 0.440 0.0 5.0 0.0 10.0 0. 0. Simplified diagram of experimental test setup is presented on Fig. Original EMTDC/PSCAD files with fully established and configured models of relays and protected cable system are available at main author on request..160 0.120 0.240 0. ● restraint charge QRES │ ● operational charge QOP 15.0 12.040 0. External Fault State at 10 ms: a) Phasor Comparison Technique. . Operational values significantly exceed restraint ones after 20 ms from fault occurrence for phasor and 15ms for charge comparison principles.. Single-Phase Intenal Fault at KYV substation Simulation test involves internal fault appearance in phase A within protected cable.0 0..0 A 6..480 0. .000 0. .120 0.320 66. 0.000 IopC_phasor IresC _phasor IopC_charge C phase (A·s) P ASE : G H _C raphs P ASE H _C : Graphs IresC _charge 40.160 0.080 0. 0.160 0.240 0.0 36. 0..200 0.0 15.0 12.0 35.0 48.0 0.040 0.0 6.200 0.240 0..50 1. cross-blocking feature is introduced. B.120 0. As the worst case scenario..0 30.0 30.50 2. This means that transformed secondary current lies within fault area and security margin is increased in corresponding phases..0 A IopA_phasor IresA_phasor 12.040 0.0 30. Hence.0 48.0 54. Two-Phase External Fault at KYV substation External fault simulation in phases A and B allows analysis on how relay computer model reacts when high currents flow through the protected cable. single Aphase fault is applied with high resistance Rfault = 20 and all shunt reactors are switched ON (lowest charging current). Computer model with established setting parameters should be able to properly detect and recognize fault state within phases.0 10..200 0..240 P HASE : Graphs _C 15.00 1.080 0.040 0..400 0. In EMTDC/PSCAD computer model.0 ● restraint phasor IRES │ ● operational phasor IOP A phase (A·s) A phase (A) a) P ASE H _A_1 : Graphs IopA_phasor IresA_phasor QopA_charge b) P ASE H _A_1 : Graphs Q resA_charge 66.50 4..0 9.120 0.0 54. transition takes place resulting in switching multiplying factors from P2 to P3 A.120 0.000 0.360 0. As a result.0 A 6. A IRES IOP 0.160 0.120 0.200 0.. It is seen that during whole simulation operational values do not exceed restraint ones.320 0..240 0.0 5.0 15..0 0. 0.0 9..320 IopB_phasor IresB_phasor IopB_charge P ASE H _B_1 : Graphs IresB_charge 66.be combined with logic gates so that tripping signal depends on the resulting signal from the inrush restraint feature.0 48.0 30. 0.0 A 20. RELAY TESTING A B phase (A·s) A 0.240 B phase (A·s) B phase (A) IV.480 .120 0..040 .0 18. ● restraint phasor IRES │ ● operational phasor IOP 5. 7. restraint values return to its normal threshold levels since transformed current lies once again within load area.50 4. cross-blocking utilizes single phase tripping and inrush activation signals as the output signals of each phase module.240 0.080 .

Differential threshold parameter Idiff> for phasor comparison was adjusted in order to obtain its critical threshold values. vol. Germany: Siemens AG. By the analysis of restraint/operational plots in EMTDC/PSCAD computer software.Energinet. FO cable DIGSI 4.dk. 335-340. IET.TABLE VII.33 Figure 10. 8 8 14 │ 14 18 │ 12 │ 18 14 14 15 │ 15 15 │ 10 │ 15 2) Sensitivity: Relay’s sensitivity analysis is critical for proper internal fault states recognition. pp.81 Advanced Transplay UDC supply Ia11 Ib11 Ic11 UDC supply Ia22 Ib22 Ic22 VALIDATION RESULTS OF SENSITIVITY FOR EMTDC/PSCAD RELAY MODEL Differential phasor Idiff> threshold for tripping Experimental Results PSCAD Simulation Results Internal fault resistance ( ) 55 70 Differential phasor Idiff> threshold for non-tripping Experimental Results PSCAD Simulation Results 4.02 3. giving overall 10 different study cases [2]. 2009. M.74 1. Ib22. Ziegler. “Protection philosophies for HVAC transmission network. For this reason. Applications of PSCAD/EMTDC Application Guide. Numerical Differential Protection: Principles and Applications.24 1. All simulation results have been successfully compliant with the expected ones. Analyzed study case results are presented on table VI. According to simulation results. 2002. 25.31 4. Ib11.03 3.1 Manual.05 3. VALIDATION RESULTS OF OPERATING SPEED FOR EMTDC/PSCAD RELAY MODEL Time interval (ms) Case description Single-phase to ground fault in the middle Single-phase to ground fault at KYV busbar Two-phase to ground fault at KYV busbar Three-phase to ground fault KYV busbar internal internal internal internal Experimental Results PSCAD Simulation Results EMTDC/PSCAD relay computer model proves to be reliable and efficient from taken simulation cases with established parameter set. 1) Operating Speed: Operating speed analysis gives idea on how fast relay is able to detect fault states. 145 210 Six current signals are sent: Ia11.makes no reaction for the same fault conditions. Aalborg. Nuremberg. These logs can then be sent to PC and read in DIGSI software for further analysis and for comparison purposes. Germany. Very high fault resistances from which relay cannot detect faults marks efficiency of specific algorithms implemented and used for measurement and comparison purposes of the obtained signals. Press.” IJEEE: Manchester Univ. Ic22 from which three enter to each relay accordingly to the side from which they were measured.S. Relays interconnected together with fibre optic cable. Sztykiel. Manitoba HVDC Research Centre Inc. which delivered all necessary technical data of the analyzed system along with Siemens SIPROTEC 4 7SD522 relays.73 1.. Winnipeg. Aalborg Univ. Germany. Phasor comparison is examined since it is more sensitive and necessary for proper fault detection.32 4. 2005. “Representation of Magnetisation Curves over a wide region using a non-integer power series.73 1.if increased of a single setting step . T. analyzed simulation cases included transmission cable’s energization and shunt reactor’s energization states. Results are listed and compared in table VII. relay model is able to accurately detect internal faults and differentiate them with mentioned other states that may be misleading. Nuremberg. Experimental Test Setup.06 3. Jul. “Field Experience Summary with a Line Differential Relay Using Complex Communication Infrastructure. Inc.. J. time interval between exceeding threshold by operational charge value QOP and phasor value IOP can be compared with the ones obtained from DIGSI logs.26 1.23 1.” M. Relay’s operating speed and sensitivity have been examined. VI. Critical values are these on which relay still detects fault and .72 1. Nuremberg. CONCLUSIONS B. 2009. ACKNOWLEDGMENT The first author gratefully acknowledges research support from the Danish TSO . internal faults with very high resistance values were analyzed. SIPROTEC 4 Differential Protection 7SD52 V4. Ic11. Results All tests from experimental analysis and EMTDC/PSCAD simulations were performed with the same setting parameter values. VII. Apart from internal and external fault states. Oct. Canada. Denmark. respond based on delivered signals with measured values and annunciation messages saved as logs..25 1. Siemens AG. VIII.. Sep. Ia22.32 4.” Siemens Energy. Rohan Lucas. thesis. Sezi at al. REFERENCES [1] [2] [3] [4] [5] [6] G. No 4. 1988. TABLE VI. .