INTELLIGENT PROTECTION RELAY SYSTEM

FOR SMART GRID

t
t
F. Kawano*, G.P. Baber*, P.G. Beaumont*, K. Fukushima , T. Miyoshi,
#
#
t
t
t
T. Shono, M. o okubo, T. Tanaka, K. Abe, S. Umeda
*

ToshibaInternational (Europe) Ltd, UK,

t TOSHIBA Corporation, Japan
# Shikoku Electric Power CO.INC,
,
Japan

Keywords: Smart Grid, protection relay, power system

characteristic, facility management, agent technology

Abstract

The authors suggest the concepts of protection relay systems

for operation within a Smart Grid and describe the results of a
prototype development. This paper introduces each of the
system characteristics that should be considered for protection
operation within Smart Grid, and the evaluation methods that
were applied under both normal and faulted conditions. The
results of a field trial in the 66kV system operated by Shikoku
Electric Power CO.,INC indicate that the proposed system
has the capability of evaluating power system characteristics
in detail on a real-time basis and contributing toward
supervision, coordination check and optimisation of setting
values. Moreover, it has been proven that an agent system can
be applied under practical conditions. Further developments
aimed at commercial operation will be considered in future.
1. Concepts of protection relay system for Smart

Grid

In recent years, trends toward electricity liberalisation and the

increased importance of environmental issues have led to
diversification of power system configurations and
characteristics, through the installation of distributed power
sources based on photovoltaic and wind power generation
technologies. Under these circumstances, there is concern
regarding the effect on the reliability of power networks with
respect to aspects such as overload, frequency variation, and
harmonic content. Against this background, there has been
increased focus on the concept of "Smart Grid", which has
the capability of collecting wide-area power system
information on a real-time basis from intelligent power grids
usingInformation and Communication Technology (lCT).

(a) Protection relay operation suitable for actual power system

characteristics
Evaluation of relay setting values can be performed by means
of supervising the operating and non-operating margins of
each relay characteristic quantitatively and detecting
variations of power system characteristics precisely on a real
time basis utilising analogue input data under normal and
faulted conditions. This approach leads to benefits such as
modifications to the setting values suitable for actual power
system characteristics, coordination check for each protection
relay element based on wide-area data, and critical point
detection for system operations.
(b)Intelligent collection of device data
Each protection relay device is provided with the
aforementioned function and carries out first-stage evaluation,
before transmitting only the data necessary for analysis.In
addition, an agent system is applied, which has the capability
of collecting protection relay data intelligently and evaluating
data from wide-area protection relays without having any
effect on protection functions. This approach can realise
updating of applications executed in protection relays without
needing to take equipment out of service.
2 Evaluation of power system characteristics
and relay setting values

Based on the concepts mentioned above, the authors have

designed and developed a prototype protection relay system.
Firstly, each of the system characteristics that should be
considered under Smart Grid and their respective evaluation
criteria will be described. Next, the evaluation methods for
both normal and faulted conditions based on examples of
system impedances will be introduced.
2.1

Although the setting values of protection relays in power

systems are currently decided based on the power system
conditions assumed in advance, instances of unnecessary or
missed operations might occur in future due to unexpected
power system phenomena or operations.In this paper, the
concepts of protection relay systems for Smart Grid are
suggested as follows:

System

characteristics

and

respective

evaluation

criteria

A key factor of the prototype system is the question of how to

capture changes in system characteristics in terms of
evaluating the margin and the coordination of relay settings
relative to the change in system characteristics. Rather than
detecting abnormal conditions directly from the change in

system characteristics, the prototype system uses evaluation

criteria related to each system characteristic within each
protection relay unit, as shown in Table 1. On detection of
abnormal conditions, the related disturbance record data can
be sent to the information management servers, and a detailed
evaluation can be executed with locus data on the
characteristic plane being calculated in the server, and finally
the results can be displayed on the client Pc. The final
confirmation of a change of system characteristics and
modification of setting values would be made by operators of
this system. In some cases the change in system
characteristics can be captured based only on the power
system quantities under normal conditions, and in other cases
only under fault conditions. The power system quantities
required for the evaluation differ for the various system
characteristics, as shown in Table 1.
NorEvaluation
System
Fault
characteristics
mal
criteria
Three-phase
Zero-sequence
...;
1
currents
unbalance
Unbalance
Segregated
...;
...;
2
between IL
transverse diff.
and 2L
current
Zero-sequence
Zero sequence
...;
...;
circulating
transverse diff.
3
current
current
Load I Fault
Load
...;t
...;*
4
impedance
conditions
Charging
...;
...;
Diff. current
5
current
The effect of
...;
...;
Diff. current
branch load or
6
source
Transient
...;t
7
Fault impedance
overreach
Overreach of
...;t
8
Fault impedance
lead phase
...;
CT saturation
Diff. current
9
T
descnbed m 2.2, descnbed m 2.3
Table 1: System characteristics and evaluation criteria

No.

2.2 Evaluation method for normal conditions

Under normal conditions, the evaluation is executed

periodically utilising power system quantities, which are
routinely measured within each protection relay unit. As an
example, an evaluation method based on the system
impedance will be introduced here. Under normal conditions,
an evaluation focused on the load impedance would be
executed so as to confirm the margin to non-operation
between load impedance and the mho characteristic (or
quadrilateral characteristic) of a distance relay element.
Hence, as shown in Figure 1, an abnormal condition could be
detected by means of whether or not the load impedance locus
enters the indicated region, which has a judgement threshold
K_ D * ZL9, providing sufficient margin to the setting of the
mho characteristic. Storage of a disturbance record would be
triggered by this evaluation result, and this recorded data

would be sent to the information management servers

simultaneously. This evaluation would be performed twice
within each protection relay unit in order to prevent incorrect
detection caused by transient phenomena.
K 0 * ZL6

ZL6
\
\

Figure 1: Evaluation characteristic under normal conditions

2.3 Evaluation method for fault conditions

Under fault conditions, each protection relay unit stores

voltage and current input data as disturbance records triggered
by the operations of UV/OC relay elements, and the
evaluation is executed utilising these data. As an example, an
evaluation method based on the fault impedance will be
introduced here. Under fault conditions, the evaluation
focused on impedance would be executed so as to confirm:
(1) the margin to operation or non-operation between fault
impedance and the mho characteristic (or a quadrilateral
characteristic) of a distance relay element; (2) the risk of
transient overreach, and; (3) the risk of overreach of the
leading phase. Therefore, as shown in Figure 2, judgement
would be made by means of whether or not the fault
impedance locus enters the region consisting of the mho and
reactance characteristics, which have threshold values K_ D' *
ZL9 and K_ XI' * Xl, so as to detect forward fault cases
which occur just inside or outside of zone 1. Since
disturbance records triggered by the operations of the UVIOC
relay elements are utilised, this evaluation could also cover
the external fault cases of relevant protection relay unit.
Hence, for example, the non-operation margin to the load
blinder element of a distance relay under external fault cases
could be evaluated properly.
K 0'* ZL6

ZL6

"

K Xl'*Xl
Xl
\

Locus

Figure 2: Evaluation characteristic under fault conditions

3 System configurations

can dispatch this application to each protection relay for

execution depending on the situation.

The prototype system is composed of information

management servers, clients and protection relays shown in
Figure 3. All devices are connected to each other via a
network over which they can send and receive relevant data.
Although there are no physical constraints on the information
management servers for practical operation, we recommend
that servers should be located in each engineering office of a
power utility in view of distributing the processing load and
maintenance activities.
"A" Office: Information

3.2 Protection relay

The protection relays implement an Agent Platform (AP),

which is a mechanism for executing applications stored in the
information management servers and processing assigned
functions. The function outline of an application which could
be processed on the AP and stored in information
management servers will be introduced below.

"B" Office: Information

A network processing module in each protection relay has

functions which provide the capability of detecting changes in
Client
management server
management server
system characteristics under both normal and fault conditions.
If this module detects a change, the related disturbance record
data stored under the conditions below is sent to the server.
Under normal conditions, the network processing module
detects changes in system characteristics based on power
system quantities measured periodically, and stores a
1 ;:
i r
II
disturbance
record. Conversely, under fault conditions, the
I
I
I
II
I I
I
disturbance
record
would be stored in case of detecting an
I I
I
: :
II
RY : I:
:
internal
or
external
fault based on operation of OC/EF relay
II
I
I
I
II
I
I I
I
elements.
Since
system
characteristics are evaluated within
I
1
"c" substation
!..."D" substation _J "E" substation _J
each protection relay in this way, the system avoids the
Figure 3: System configuration
unnecessary transmission of data from each protection relay
to the information management servers, and as a result, an
Taking into consideration the need for application updates in
intelligent system is realized.
accordance with changing requirements for the system in the
future, the system consists of agent technology based
In addition, each protection relay will also notify status
configurations, in order to provide sufficient flexibility.
change information to the information management server
Therefore, any of the applications executed in each protection
when the OC/EF relay elements operate. Thus, the system
relay can be updated online without needing to take
operators can collect the desired disturbance record data from
equipment out of service. In addition, this system allows
the related protection relay unit manually and analyse the
connection of multiple vendors' intelligent electronic devices
system phenomena detail in the server.
(lEDs) such as protection relays, network terminals and smart
meters. This paper will explain in detail the allotment of
3.3 Client
functions to information management servers, protection
relays and client PC.
The client in this system provides Human MachineInterface
---------

__ -

-------

------------

----

___

3.1 Information management server

The information management server has both data server

functions and application server functions. The former are
functions which collect, store and analyse the necessary data
from each protection relay, while the latter are functions
which store the application software executed in each
protection relay and manage the execution status in each as
shown below.
(a) Data server functions
These functions collect disturbance record data from each
protection relay, calculate locus coordinate data and detect
changes in system characteristics precisely based on each
evaluation method by utilising disturbance record data
collected from each protection relay.
(b) Application server functions
The application software executed in each protection relay is
stored in the information management servers. The servers

(HMI) functions and implements functions for displaying

information as well as maintenance management functions.

(a)Information display functions

This is the function which displays data and analysis results
stored in the information management server. The client can
display current and voltage waveforms, evaluation results of
system characteristics and protection relay settings.
(b) Maintenance management functions
The evaluation parameters of the application executed in the
information management servers and protection relays can be
set in the client. A screen example of a client is shown in
Figure 4.

expressions (I) and (2) respectively, and the evaluation

characteristics for impedance are shown in Figures 1 and 2.

Ia(1L)
I
-Ia(2L) I > K_ID *ID
Io(1L)
I
-Io(2L) I > K_IOD *IOD

Key:
ID:

..

IOD:
KID:
KIOD:

Setting value for segregated transverse

differential current relay element [A]
Setting value for zero-sequence transverse
differential current relay element [A]
Evaluation threshold for segregated transverse
differential current relay element [%]
Evaluation threshold for zero-sequence
transverse differential current relay element [%]

The field trial was established with high sensitivity evaluation

thresholds compared with relevant relay settings as shown in
Table 2.

Figure 4: Example of Client Display

4 Field trial and evaluation

In order to demonstrate the capability of this system, a field

trial was established in the 66kV system operated by Shikoku
Electric Power CO.,INC from June 2008 to February 2009.
The system configuration for the field trial is shown in Figure
5. An information management server and a client were
located within the head office, three protection relays (RYI,
RY2 and RY3) were situated within two substations, and a
dedicated network was provided for maintenance. Although
no internal faults occurred during the operation of the trial
system, the results of the evaluation of power transient
fluctuations utilising wide-area data during system operations,
e.g. disconnecting one circuit, lL shown in Figure 5, of a
double-circuit parallel line will be introduced below.
I' Head office
I
I
I
I
I
I
I

--

"A"

Substation

2L
66kV double
circuit parallel
line
L________________

(1)
(2)

____________

Figure 5: System configuration for the field trial

In this field trial, zero-sequence and segregated transverse

differential current as well as impedance as shown in Figure 1
were applied as the evaluation criteria under normal and fault
conditions. The evaluation equations for zero-sequence and
segregated transverse current are shown by the following

Evaluation criteria

Evaluation threshold [%]

Normal Fault
Name
cases
cases

Zero-sequence
KIOD
60
transverse diff. current
Segregated transverse
KID
60
diff. current
Mho
K D
140
Impedance
Reactance K Xl
----Table 2: EvaluatIOn thresholds ill the field trial

80
80
120
120

The lL and 2L current waveforms of RYl, which were

collected in the information management server on
disconnection of lL of the double-circuit parallel line, are
shown in Figures 6 and 7. These waveforms conform to the
system phenomenon because the lL currents became "0" and
the 2L currents doubled just after this system operation. In
addition, it could be confirmed simultaneously that the
disturbance record data stored by the trigger of the OCIEF
relay operation were properly sent to the information
management server based on the evaluation executed in RYI.
The loci on the relay characteristics of the segregated and
zero-sequence transverse differential currents based on the
disturbance record data are shown in Figures 8 and 9.
In Figure 8, although some transverse currents for each phase
are generated, it can easily be recognised that none of the loci
exceed the threshold. On the other hand in Figure 9, in
accordance with the evaluation algorithm, the change in
system characteristics can be detected correctly by means of
evaluating the zero-sequence transverse currents, which are
generated transiently just after the lL disconnection.
Through the field trial, it was demonstrated that this
intelligent system has the capability of monitoring system
phenomena appropriately and evaluating the validity of relay
setting values based on the system phenomena which would
be faced under Smart Grids. A total of 31 notifications of
changes in system characteristics were issued during the field

trial, and all necessary data was correctly collected at the

information management server. This indicates that the AP in
the protection relays, the agent management mechanism and
the agent application system can meet the demands of
practical conditions.
2. 0

BJa

1.5

1.0
[A]

0.5

0.0

-0.5

-1.0

-Ib

1. "" \
- -

' \ 83 y

166

V '-i

-1.5

threshold

-1.0

Figure 9: Locus of zero-sequence transverse differential

current on disconnection of lL of the double-circuit parallel
line

,.-------,

_ --tt-1r
r :
0
+

f---"

IU -I-0.5 I vr:t vit

-1.0

threshold
- Relay setting

5 Conclusion

: ,-_C

O. 0

[A]
'----'*H-+t--L-----'
0.5
-1.0
1.0

[ms]

Figure 6: Waveforms of currents (IL) on disconnection of lL

of the double-circuit parallel line

[A]

0.5

- Evaluation

Ie

-2. 0

0.5

-+-10(1 L)-IO(2L)

-10

-1.5

2. 0

1.0

f3n

1-\-IrI- -/-HII_M=+#*_+l'*_+t_#+i

11, 1'\ 'iJf

1-------++ -

-:g
-

-t-

[ms]

56

-2.0

Figure 7: Waveforms of currents (2L) on disconnection of lL

of the double-circuit parallel line
2.5

__ la(1 L)-la(2L)
__ lb(1 L)-lb(2L)

Ic(2L)-lc(2L)
- Evaluation
threshold

[ -" ] - Relay setting

A
,+--+-=-.-#X''--L-_-,---+--+threshold
-2.
2.!"

-2.5

Figure 8: Loci of segregated transverse differential currents

on disconnection of lL of the double-circuit parallel line

The authors have proposed the concepts of intelligent

protection relay systems for Smart Grid and have proceeded
with a prototype development based on these concepts. The
results of the field trial of this system demonstrate that the
proposed system has the capability of evaluating power
system characteristics in detail on a real-time basis and
contributing toward supervision, coordination check and
optimisation of setting values. Moreover, it has been proven
that an agent system can be applied under practical conditions.
Further developments aimed at commercial operation will be
considered in the future.
Acknowledgements

The authors wish to thank the staff of Shikoku Electric Power

CO.INC
,
for their support before and during the course of the
site trial.
References

[1] S. Umeda, A. Matsuno, K. Abe, M. Hirao, M. Takahashi,

K. Fukushima, F. Kawano, M. Ookubo, T. Miyoshi. "New
Relay System Supported to High Confidence Operating
on Real-time", The Papers of Technical Meeting on
Power Protective Relay, PPR-09, pp. 45-50, (2009).
[2] S. Umeda, A. Matsuno, K. Abe, F. Kawano, M, Ookubo,
T. Miyoshi, K. Fukushima. ''New Relay System
Supported to High Confidence Operating on Realtime",
The 2009 Annual Meeting of the Institute of Electrical
Engineers of Japan, 6-320 (2009).
[3] S. Umeda, A. Matsuno, K. Abe, F. Kawano, K.
Fukushima, T. Miyoshi, M. Ookubo. "Field test result of
New Relay System Supported to High Confidence
Operating on Realtime", Proceedings of the Twentieth
Annual Conference of Power & Energy Society, 188, pp.
13-14, (2009).