Gridtech Subs Auto Smart

SESSION
4
Emerging Trends in
Substation Technology - I
5th International Exhibition & Conference, 2015

April 8-10
2015
New Technologies in T & D, Renewable Energy Integration,

Smart Grid, Energy Efficiency and Communication
212

April 8-10
2015

Special Protection Scheme for Bus Reactor in

Transfer Bus Coupler Bay
N.M. Sheth
S.K. Jadav
B.J. Patel
Gujarat Energy Transmission Corporation, India
SUMMARY
GETCO network is expanding rapidly to meet the
generation as well as demand requirements in the
state. Also due to non availability of gas,
government policies in the benefit of RE
generation and implementation of merit order
dispatch, power flow pattern has changed
completely in last few years and load pockets have
become generation pocket also. Earlier
generation pockets in Gujarat were in southern
region and power was flowing towards central,
western and northern region. But, now power is
flowing from western region to central, southern
as well as northern parts of the Gujarat.
Moreover, western and northern region is having
predominant agriculture load. Hence, during nonagriculture season load consumption reduces
considerably. In addition to that, transmission lines
connecting these regions with central Gujarat are
quite long. These long lines are loaded below SIL
resulting in generation of excess reactive power
causing high voltage at 400KV Buses in these
areas. Because of Over Voltage, many 400KV
Lines are required to be switched off to maintain
voltage within acceptable limits. This, in turn, is
jeopardizing grid availability and security.
To overcome Over Voltage, GETCO found the
need of Reactors at 400KV strategic location (7
Places) based on system study. But, out of seven,
four locations have the space constraint for
providing Reactor bay. After brain storming, it
was decided to utilize TBC bay as Bus Reactor
bay by providing Bus reactor in Transfer bus
coupler bay. The main consideration behind this
decision was utilization of TBC bay on very few
occasions in a year. On finalization of this
modality, there are challenges regarding (i) Bus

Reactor arrangement in TBC Bay, (ii) Control &
Interlocking philosophies, (iii) Protection
philosophies, (iv) Whether to utilize TBC Bay as
TBC as well as Bus Reactor or either of the two.
Challenges become harder as each substation has
different Make and Type of TBC Panels, Different
TBC interface schemes and Different Bus bar
schemes. Also there are challenges regarding
which type of panel is to be considered for Reactor.
The above scheme has really helped GETCO in
mitigating Over Voltage problems substantiously
and there-by strengthening security of grid. This
paper shares the knowledge of implementing the
scheme, issues & challenges faced and benefits
achieved.
Keywords: TBC, BCT, RADSS, PS class
1.
INTRODUCTION
In recent years, power flow pattern has changed in

Gujarat state because of load growth in specific
pockets and techno-economic considerations like:
(a) Shutting down of Gas based power plants due to
non-availability of Gas (b) Government policies
promoting RE generation (c) Convergence of Load
into Generation pockets and (d) Implementation of
merit order dispatch.
Moreover, western and northern region is having
predominant agriculture load and during nonagriculture season, load consumption reduces
considerably. Moreover, transmission lines
connecting these regions with Central Gujarat are
quite long. These long lines are loaded below SIL
resulting in generation of excess reactive power
causing high voltage at 400 kV Buses in these areas.
Hence GETCO network started facing over voltage
213

April 8-10
2015

in non-agriculture period. Details are mentioned in

Table 1.
Table 1
are required to be integrated precisely without

making any harm to existing TBC bay functionality.
(i) Switch Yard Arrangement
400KV
Substation
Month wise Maximum Voltage (KV)

Apr
May
Jun
Jul
Aug
2011
2011
2011 2011
2011
Jetpur
432
429
429
430
431
Bus Reactor is provided in front of Transfer Bus

Coupler bay (Fig.1). Now there were two
options for connecting Reactor with TBC Bay
as under.
Soja
421
418
417
417
419
Connection through Transfer Bus
Chorania
428
423
416
419
419
Kansari
420
415
414
423
423
Connection through bottom of Pantograph

isolator of transfer bus
Amreli
418
418
416
417
418
In case of first option, issues would be as under:-
Vadavi
426
425
425
424
423
(a) Physical connectivity with Transfer Bus
Hadala
430
427
428
431
431
(b) Bus Reactor will always remain connected

to Transfer Bus.
To over come this issue, 400KV Bus Reactor is

proposed at 7 strategic locations based on system
study. But due to space constraint, it is decided to
provide Bus Reactor in Transfer Bus Coupler bay at
four out of seven locations after reviewing historical
data of utilization of TBC Bay.
This led to new challenges for Control, Interlocking
as well as Protection philosophies including physical
connection of Reactor Bay with TBC Bay.
All challenges are addressed iteratively and after
having lot of exercise, project specific solutions are
derived and implemented for each substation.
2.
PHILOSOPHIES ADOPTED
In order to consider Bus Reactor in TBC Bay,

philosophies of TBC Bay as well as Bus Reactor Bay
(c) Transfer Bus will unnecessarily remain

charged when Bus Reactor is in service
though no bay is transferred on Transfer
Bus.
(d) Additional CT is required for balancing
Zone-3 current of TBC CT in Bus bar
protection when Bus Reactor is in service.
As there is no CT core in Reactor in phase
side matching with TBC bay CT core.
Hence this option is not considered.
Whereas second option looked more viable as
clear demarcation of TBC functionality either
as Transfer Bus Coupler bay or as Bus Reactor
bay can be achieved easily. Accordingly second
option is considered as shown in Figs.1&2.
Fig. 1: 400 kV Transfer BUS Coupler Bay with 400 kV BUS Reactor
214

April 8-10
2015

through Bottom of TBC bay Transfer bus

Pantograph isolator.
z
Separate isolator with earth switch is to be

provided for Reactor.
No separate CTs are to be provided for Bus

Reactor as Reactor Bushing CTs and spare
core of TBC bay CT will suffice the
requirement of metering as well as
protection of Reactor bay. (mentioned in
protection philosophy).
(iii) Control Philosophy

z
No separate control for Reactor Bay is

required as it is a part of TBC Bay. TBC
Bay control philosophy is applicable for
Reactor also except additional controls of
Reactor Isolator & Earth switch.
Annunciator in existing TBC Control panel

is to be retrofitted with higher window
annunciator to include reactor bay troubles
/ events.
(iv) Interlocking Philosophy

Main task was to decide and finalize interlocking
scheme such that all operational requirements
get fulfilled.
Interlocking philosophy derived is as under:
z
Reactor Isolator cannot be closed until

Transfer Bus Isolator is open.
Transfer Bus Isolator of TBC cannot be

closed until Reactor Isolator is open.
Fig. 2
z
(ii) Bay Philosophy

With the arrangement finalized as above Bay
equipment philosophy is considered as under:z
Bus Reactor is to be connected to TBC Bay
215
All bay Transfer Bus Isolator control

scheme (i.e. ladder network) DC supply
cannot be extended until Reactor Isolator
is open.
Circuit Breaker of TBC Bay can also be
Closed and Opened when Reactor Isolator
is closed (existing logic is only when
Transfer Bus Isolator is closed).
TBC bay PT selection scheme is also
required to be operated when TBC is
working as Bus Reactor Bay. Accordingly,
Reactor Isolator close condition is also
included in parallel to Transfer Bus Isolator
close condition.

April 8-10
2015

(v) Bus bar protection scheme Philosophy

z
TBC Bay CT core is routed to Z-3 of RADSS

Bus bar scheme without CT switching.
Hence it is required to be blocked when
TBC bay is working as Reactor Bay to avoid
Z-3 operation.
Accordingly, CT core is blocked through
Reactor Isolator contacts (i.e. core shorted
when Reactor Isolator is closed and routed
to Z-3 module only when Reactor Isolator
is open.
Reactor Differential Protection: TBC Bay

spare PS class core and Reactor Bushing
CT Neutral side core (PS class) are utilized.
(500/1 Ratio)
Reactor Back-up Impedance Protection:

Reactor Bushing CT Core is utilized (PS
class 200/1 Ratio).
Reactor REF Protection:Reactor Bus side and Neutral side Bushing

CT cores (PS class, 200/1 Ratio) are
utilized.
(vi) CT Core utilization for Metering and

Protection (Fig. 3)
z
Reactor Bus bar Protection:No need of separate CT core as it is not a

separate Bay. TBC bay CT cores routing to
Z-1/Z-2 of Bus bar scheme is applicable to
Bus Reactor in TBC Bay also.
Metering: Reactor Bushing CT core is

utilized (0.5 class, 200/1 Ratio).
(Reactor Bushing CT cores towards Bus side

has only 200/1 Ratio hence it can not be
utilized for Bus Bar protection)
3.
FINALIZATION OF PANEL
REQUIREMENT
As this is not being a regular Reactor bay, scheme

design is finalized based on philosophy and
considerations mentioned in previous section and
following points are concluded:
z
Separate Control Panel for Bus Reactor is not

required. But, separate metering, Reactor isolator
control, interlocking and Reactor OTI, WTI
measurements are required.
All Reactor protection and Reactor Trouble

relays are required.
Trip circuit supervision relay, CVT selection

relays etc. are not required as existing TBC Panel
relays are to be utilized for this.
Fig. 3
216
Separate energy meter is required for Bus

Reactor as TBC Panel does not have energy
meter.
Modifications in existing TBC Panel are required
for interface with Bus Reactor Panel.

April 8-10
2015

The above requirements are summarized in Table 2.

Table 2
Sl. No.
1
4.
Items considered
Relay Panel for Bus reactor (with Energy meter, Indicating meters, WTI, Isolator control switch,
necessary auxiliary / multiplication relays for Isolator & Earth switch etc.) including,
z
Related interface wiring & modification in existing TBC panel with required accessoires
and complete scheme testing.
z
Supply of 36 Window annunciator and retrofitting of same in existing TBC Panel with
related wiring and accessories.
z
Modification in Mimic Diagram of existing TBC panel as per revised bay configuration
along with required accessories and wiring.
PREPARATION OF OPERATION
PHILOSOPHY
Reactor isolator is open.
It is necessary to prepare operation philosophy for

easy understanding to operating staff to avoid any
mal operation. Accordingly; operation philosophy
prepared as under and conveyed to respective
substation.
z
Bus Reactor has its own Isolator with E/S.
Remaining equipments of TBC bay shall be

utilized for Bus Reactor.
TBC Panel scheme such as CB Close and Open,

CVT selection, LBB Positive extension etc.
works only when 89C is closed. Now with Bus
Reactor in TBC all these functionalities are also
be possible when Reactor Isolator is closed.
When TBC Bay is to be utilized as Bus Reactor

bay; Zone-3 of Bus bar scheme shall be kept
OFF along with following modifications
Bus Reactor is connected to Transfer Bus

Coupler bay through Transfer Bus Coupler bay
Pantograph Isolator.
1.
Separate Reactor protection panel is provided

for Bus Reactor whereas control will be through
TBC Control panel with required interface as
per respective substations.
Scheme will be such that; TBC Bay can either
be utilized as TBC Bay or as Bus Reactor Bay.
Both functions simultaneously will not be
possible.
2.
No any other bay should be transferred on TBC

when is it being utilized as Bus Reactor Bay
Reactor Isolator can be closed only when
Transfer Bus Isolator of TBC bay and TBC
Breaker is in open condition (also Line side E/S
of breaker if provided).
TBC Bay Transfer Bus Isolator can be closed
only when Reactor Isolator is open.
Other bays Transfer Bus Isolator supply through
ladder network can be extended only when
In case of RADSS Bus bar scheme; TBC CT

core directly routing to Zone-3 shall be
routed through 89CX such that;
z
When 89C is closed (i.e. TBC bay is

being utilized as Transfer bay) CT
should be routed for Zone-3
measurement.
When 89C is open (i.e. TBC Bay is

working as Reactor bay) CT should be
shorted through 89CX contacts.
For Numerical Bus bar protection scheme;

89R status inputs shall be routed to TBC
Bay unit and logic shall be prepared such
that Zone-3 measurement shall be blocked
when Reactor isolator is closed.
Utilization of CT core for protection & metering

is as under.
Metering:
Reactor BCT Phase side CT (200/1 ratio)
Differential Protection:
TBC Bay PS class CT core & Reactor Neutral side CT
core (500/1 ratio).
217

April 8-10
2015

REF Protection:Between Rector BCT (200/1) HV & Neutral side CT
out workable solution through innovative design

thinking. GETCO has similar issue of space constraint
for providing Bus Reactors to control Over Voltage.
Back-Up impedance protection:Reactor BCT Phase side CT (200/1 ratio)
Out of the seven, four locations were having space

constraint. TBC bay was utilized for the new Reactor
and whole scheme was commissioned with minimum
time and cost along with redundancy and better
flexibility.
5.
BENEFITS ACHIEVED
Saving in bay equipment cost.
Space optimization in switch yard.
Space optimization in control room.
No augmentation of Bus bar protection scheme,

as it is achieved with modifications of TBC
relays.
Reduced quantum
commissioning work.
Saving of constructional cost (electrical as well

as civil).
Reduced cabling as new cables from switchyard

to C/R are required to be laid for Reactor and
its Isolator only. For remaining scheme
interface, cabling is between TBC panel and
Reactor panel within Control Room.
6.
LIMITATION & RISK FACTORS
Situation may arise when there is need to

Transfer bay on TBC as well as to put Reactor
in service. Under such circumstances, either of
them is to be selected on the basis of priority.
Scheme requires extra care during testing as well

as O&M.
Mal-operation of Reactor trouble relay may

cause unnecessary tripping of TBC Bay and
interruption to transferred bay.
When TBC Bay works as Reactor bay, Zone

covered by Bus bar protection is reduced. Any
fault beyond TBC CT is covered by Reactor
Differential Relay instead of Bus bar protection.
of
erection
and
(This can be overcome by providing required

Bushing CT in Bus Reactor.)
7.
CONCLUSION
Transmission network and substations have grown

gradually with increase in power demand and there
is hardly any space left over in substation to upgrade/
accommodate new bays and support systems like
Reactor, SVC, and STATCOM. Utilities have to find
But, there will be certain limitations so this type of

scheme should not be considered as regular scheme.
Education and clear operation instructions are
necessary.
BIOGRAPHICAL DETAILS OF THE

AUTHORS
N. M. Sheth Deputy Engineer: Obtained Graduation
in Electrical Engineering from Saurashtra University
Rajkot; and Qualification of Certified Project
Management Associate (Project Management LevelD, National Ranker) from International Project
Management Association (IPMA). Working in
GETCO since 1994. Experience in the field of
Substation Operation & Maintenance as well as
Commissioning; Protection scheme design and
commissioning. Presently working in Engineering
department and responsible for Design & Engineering
of Control, Protection, Automation, schemes &
Philosophies and Substation secondary engineering.
S. K. Jadav Junior Engineer: Obtained Bachelor in
Electrical Engineering from Saurashtra University,
Rajkot and Masters in Electrical Power Engineering
from Maharaja Sayajirao University, Baroda. Working
in GETCO since 2009 and responsible for Design &
Engineering of Control, Protection, Automation,
Schemes & Philosophies and Substation secondary
engineering.
B. J. Patel Junior Engineer: Obtained Bachelor in
Electrical Engineering from Saurashtra University.
Joined Rajashree Polyfil (A division of Century Enka
Ltd.) as a Trainee Engineer in year 2004. He also
served in Suzlon Infrastructure Ltd, SEZ, Vaododara
up to year 2011 and presently working in GETCO
since 2011. Experience in the field of Substation
Operation & Maintenance as well as Commissioning;
and responsible for Design & Engineering of Control,
Protection, Automation, Schemes & Philosophies and
Substation secondary engineering.
218

April 8-10
2015

Degree of Oil Refining and its Effect on

Insulating Paper Degradation
I. Crusell
L. Bergeld
B. Pahlavanpour
Nynas AB, Sweden
SUMMARY
Furanic compound is generic name of chemical
compounds produced from degradation of
insulating paper in oil filled transformers. These
are polar compounds which are, up to certain
extend, soluble in transformer oil. By
measurement of Fufurals in oil it is considered to
be possible to estimate rate of aging of a
transformer. Water, acids and carbon monoxide
are also byproducts of paper aging and it has been
established that oil deterioration and acid
formation also have effect on degree of
polymerization of insulating paper.
In a laboratory study, of which this paper is based
on, Kraft paper was aged together with copper in
different dielectric fluids, (including a vegetable
ester, a hydrocarbon fluid based on predominantly
noncyclic hydrocarbon chains and two different
naphthenic oils). After conducted ageing water
content, furfural content, and acid number was
measured in both paper and fluids. It was
discovered that fluids with high content of
degradation products in oil not necessary
contained the most amounts of degradation
products in paper. Hence the distribution of
furfural/water/acids between paper and oil is
different for different oils.
The estimation on degree of paper degradation
based on paper degradation by products in oil is
complicated and therefor need more study.
Keywords: Mineral Transformer oil, insulating paper
aging, other insulating liquids, Fufural formation
1.
INTRODUCTION
It is not an easy decision to choose the insulating
fluid for a transformer since there are many

parameters to take into consideration, such as
environmental effects, price and function. Price is
handled during the sales process and is not a topic
for this paper. Environmental effects is a very
complex topic since any comparison of products to
determine prefer ability must assess all the relevant
environmental impacts across the full product life
cycle, including production, refining, duration of use,
emissions from transportation, recycling etc. to make
sure that the problem is not simply shifted elsewhere.
It is complex and at the same time an important topic
that needs more attention. But the aim here is to
compare mineral oil and ester fluids with respect to
function inside the transformer, focus will be on
essential properties such as oxidation stability, fire
point, DGA and fluid-cellulose interaction.
1.1 Transformer Fluids

The transformer fluid is a vital part of a transformer
as it serves several functions such as being a part of
the electrical insulation, cooling of the windings and
core, and functions as a carrier of information. There
are different alternatives on the market such as
mineral oil, synthetic hydrocarbons, natural/
synthetic ester and silicone oils, with varying
properties. This paper will focus on mineral oil and
ester fluids.
1.2 Mineral Oil

Mineral oil is the most commonly used insulating
liquid for transformer applications [1]. They consist
of hydrocarbons organised in various molecular
structures, generally classified into three main types:
paraffins, naphthenes and aromatics (Figure 1).
Mineral oils used for insulating purpose often
contain high percentages of naphthenic molecules.
219

April 8-10
2015

Generally, an oil with naphthenic characteristics

provides good low temperature properties, a high
solvent power for polar substances and lowd
viscosity index. Mineral transformer oil contains
inhibitors which delay the aging of the oil. These
inhibitors might be natural, as occur in uninhibited
mineral oils, or synthetic and added, as in inhibited
oils.
Fig. 3: Molecular structure of a synthetic ester
Fig. 1: (a) iso paraffinic, (b) normal paraffinic, (c) naphthenic,

(d) aromatic
1.3 Esters
An ester (Figure 2) is either present in the raw
material, natural ester, or formed from the reaction
of an alcohol and a fatty acid, synthetic ester.
Fig. 4: Molecular structure of a natural ester
rather than refined from petroleum base stocks or

synthesized from organic precursors. With a high
percentage of unsaturated acids the viscosity is lower,
the oxidation stability on the other hand is suffering
with a high percentage of unsaturated acids [2].
1.4 International IEC Specifications
Fig. 2: Molecular structure of an ester
Esters in general are used in application where fire

safety is of primary concern. Both synthetic and
natural esters have a high fire and flash point, see
the specification requirement in Table 1 below.
Additionally the synthetic esters show a good
oxidation and thermal stability [1]. Typical structures
of natural and synthetic esters can be seen in Figure
3 and Figure 4.
Natural esters differ from mineral oil or other fluids
in that it is an agricultural product derived from
vegetable oils (soybeans, sunflower or rapeseed)
Mineral oils have been used for a long time and this
has given rise to many different national and
international specifications and standards, one of the
most common one is IEC 60296 [3]. Today there are
also specifications for both synthetic and natural
esters within IEC, IEC 61099 [4] and IEC 62770 [5]
respectively. There are differences between these
specifications when it comes to required tests,
methods and specified limits. Table 1 shows flash
point and oxidation stability from these
specifications, which will be related to in the coming
discussion.
220

April 8-10
2015

Table 1: Specifications for different transformer fluids

Flash point C
IEC 61099
IEC 62770
135
250
250
IEC 61125C, 164/

332 or 500h
IEC 61125C, 164h
IEC 61125C, 48h
min
Oxidation stability
2.
IEC 60296
Total acidity, mg KOH/g
max
1.2 /0.3
0.3
0.6
Sludge, wt%
max
0.8 /0.05
0.01
TEST PARAMETERS AND METHODS
2.1 Oxidation Stability

The oxidation stability will indicate the life time of
the fluid and the obvious target is to have a high
oxidation stability meaning a long life time for the
fluid. The base of the discussion about oxidation
stability will be the different specification
requirements for each fluid followed by an example
from an existing study [6].
2.2 Fire Point

A high fire point and flash point is important from a
safety perspective, specifically in certain applications
such as indoor areas, mines and ships. There are
different requirements in each specification which
will be highlighted in this paper.
2.3 DGA
Insulating material within electrical equipment break
down with time and this generate gases within the
unit. The distribution of these gases can also be
related to abnormal conditions such as electrical or
thermal faults and the rate of gas generation can
indicate the severity of the fault. Dissolved gas
analysis can therefore be used as a method to
evaluate the condition of mainly the transformer but
it can also give indications on the status of the
insulating material.
DGA diagnostics for mineral oils have a long history
and are today well established. When it comes to
the relatively new insulating liquids such as esters
many aspects of the thermal and electrical design of
transformers have to be considered. Additionally the
diagnostic and condition monitoring methods,
including DGA, have to be adjusted. The DGA results
can be analyzed using various existing interpretation
methods such as the Duval triangle method and
Rogers. These methods apply to transformers filled

with mineral oil but not directly to equipment filled
with non-mineral oils and adjustments are needed
[7]
. The aim in this paper is to highlight differences in
the interpretation guides for different fluids, including
mineral oil, synthetic ester, and natural ester, and
the reference here would be the Duval triangle
method.
2.4 Fluid-cellulose Insulating System

It is well known that moisture content has a large
impact on the aging rate of the insulation system in
transformers. Several significant experimental studies
have been made and show that water is one of the
most negative factors affecting the life time of oilpaper insulation. It can reduce the degree of
polymerization [8] [9] and lower the mechanical
strength of the cellulose [10]. Earlier studies also show
that an increase of water content in the insulation
may have an impact on the electrical properties of
the insulation [11].
Cellulose seems to be preserved better in ester liquids
compared to mineral oils. A common theory and
probable explanation to this is that an ester can
dissolve more water due to a higher polarity and
therefore extract it from the paper and reduce the
degradation thereof. However there is less
information available on how the properties of the
dielectric fluids are affected when aged together with
cellulose and how that will affect the overall
transformer operation. This study both look at the
properties of the paper but mainly the properties of
the fluid, such as dielectrical dissipation factor (DDF),
acid number and water content.
This was an in-house study performed by Nynas with
the following setup. Sealed glass flasks containing
copper, insulating paper with high initial water
content and a dielectric fluid were aged at 120C in
221

April 8-10
2015

a heating oven for 17 weeks. The study was

performed using cellulose of a high initial water
content to accelerate the aging. Changes in different
properties of the liquid were observed for the period
of aging. The effects on DDF, acid number, copper
content, and water content of the fluid were measured
during the experiment. Also the acidity and water
content of the paper were monitored during the
study. The dielectric fluids used during the
experiment were a natural ester and an uninhibited
mineral oil complying with IEC 60296 [3].
3.
RESULTS AND DISCUSSION
3.1 Oxidation Stability

As can be seen in Table 1 the specifications for the
respective fluid have different limits for oxidation
stability. For mineral oils you have different
requirements depending on the quality of the fluid.
Synthetic esters need to fulfill the lowest requirement
for mineral oil, but it is commonly known that these
fluids perform well when it comes to oxidation
stability. The test set up for natural esters are the
same as for the mineral oil and the synthetic ester
but the test duration is 48 hours, to be compared
with 164 hours which would be the lowest
requirement for the others. Due to these differences
it is difficult to compare the different fluids on a
specification level. But the duration of 48 hours
suggests that the oxidation stability for natural esters
might not be as good as for the other alternatives.
One example is the study 6 as can be seen in Figure
5. Both the mineral oil and the synthetic ester show
good oxidation stability according to IEC 61125C
(164 hours) which is the test indicated in the
specifications [3, 4, 5]. Three different natural esters
Fig. 5: Oxidation for 5 insulating fluids
have here been evaluated and as the figure show

there is a big variation among the samples. But as
expected, and what will also be showed in the coming
study, natural esters have a much weaker oxidation
stability compared to mineral oil and synthetic esters.
3.2 Fire Point

As mentioned a high fire point is important from a
safety point of view. Ester fluids are often preferred
when looking at specific applications where the
safety is critical. The limits of flash point stated in
each specification can be seen in Figure 6. As
indicated the requirements for ester fluids is more
stringent compared with the one for mineral oils.
The high fire and flash point would be one of the
main benefits with the ester fluids. How relevant
250C is when it comes to a sever fault in a
transformer could be a topic for another paper.
Fig. 6: Flash point
3.3 DGA
Depending on what insulating material that is used
there will be differences in the gas formation and
relation to a certain fault and the interpretation of
the DGA results need to take this into consideration.
The method to evaluate gas formation in mineral oil
is well established and it has recently been studied
if new interpretation methods are needed for ester
fluids or if existing methods can be slightly adjusted
to serve the purpose. This study [7] is comparing 4
fluids: 1 mineral oil, 1 synthetic ester and 2 natural
esters. A comparison of differences in gas formation
between the 3 non-mineral oil fluids with the mineral
oil have been used to adjust the zones in the Duval
triangle leading to one triangle per fluid as can be
seen in Figure 7.
222

April 8-10
2015

for the mineral oil the water content slowly increased.

The decrease of water in the natural ester is thought
to be due to hydrolysis of the ester to the
corresponding acid, which is a reaction consuming
water. It was also expected to see that the acidity
and DDF increased more for the natural ester
compared to the mineral oil due to the different
oxidation stability behaviours.
In the paper-fluid-copper system, the water content
of the fluids was observed to increase during the
first week followed by a decrease to finally level off
in a low water content in both fluid (Figure 8a) and
paper (Figure 8b). This initial increase is probably
due to a shift of water from paper to the fluid for the
system to reach equilibrium.
Fig. 7: Duval triangle for transformer filled with a) mineral oil
(Duval triangle 1), (b) synthetic ester (Duval triangle 3), (c)
natural ester (Duval triangle 3), (d) natural ester (Duval triangle 3)
When comparing these triangles there are differences

mainly in the boundaries for the zones D1/D2, T1/
T2, and T2/T3. It can also be seen that different types
of natural esters have different gas formation
patterns. A variation in chemical structure of these
fluids would be a possible explanation for the gas
formation behavior 7.
The awareness of these differences is important to
be able to make correct actions from a DGA
measurement. It is also important to have in mind if
and when there is a need to top up an ester filled
transformer with new ester. If the esters are not from
same origin there might be a change in the Duval
triangle that needs to be considered.
These triangles have been created based on tests
simulated in a laboratory and they might have to be
modified when actual cases occur [7].
Fig. 8: (a) water in fluid over 17 weeks, (b) water in paper

over 17 weeks
3.4 Fluid-cellulose Insulating System

To verify the test setup blank samples without paper
and copper was first evaluated. As expected the initial
water content of the natural ester was higher (about
10 times) than that in mineral oil. That is because
the water saturation of the ester is higher than that
of mineral oil. It can also take in water in chemically
bounded form while in mineral oil water is only
dissolved [12]. The water content of the natural ester
decreased during the time of the experiment whereas
The decreasing water content could be an indication

of that water is used in the mechanism of paper
degradation and then mainly hydrolysis by cleavage
of the inter-unit linkages. The acidity of the paper,
in the case with natural ester, was lower than for the
mineral oil, this could indicate that the decrease of
water in the ester-fluid system could be an effect of
the consumption of water through hydrolysis of the
ester rather than the cellulose.
223

April 8-10
2015

In Figure 9 it can be seen that the natural ester yields

high number of total acids when aged compared to
the mineral oil that show a more modest change.
The acidity of the fluids corresponds well with
insulation fluids expected stability towards
oxidation. The fluids were in this study aged in a
closed environment with a low oxygen content,
giving different test conditions compared to the IEC
oxidation stability test (IEC 61125C) where there is
a full access of oxygen. This indicates that the
degradation of esters is highly dependent on both
oxygen and water content present in the system. One
explanation to the higher acidity for the natural ester
could be a hydrolysis reaction of the natural ester
resulting in increased number of acids. The acids
formed in the natural ester via hydrolysis are of high
molecular weight. In comparison to low molecular
weight acids, acids of high molecular weight are not
thought to have any significant effect on the acid
catalysed degradation of paper [11,13]. This study
indicates that there is no correlation between the
acid numbers in the fluid compared with the acid
number in paper. Hence a high acid number in the
fluid will not necessary be equivalent a high acid
number within the paper.
Fig. 10: DDF in natural ester and uninhibited mineral oil
copper content, since a high copper content would

affect the insulating properties. Both un-aged samples
and samples of fluids aged for 17 weeks were sent
to an external laboratory. The results are reported in
Table 2 below.
Table 2: Specifications for different
transformer fluids
Initial copper Copper content
content in
in fluid after
fluid (ppm) 17 weeks (ppm)
Natural Ester,
Kraft paper
<0,1
17
Uninhibited mineral
oil, Kraft paper
<0,1
0,3
The copper content in the natural ester after testing

was very high compared to the content in the mineral
oil. This indicates that the natural ester can have a
corroding effect on the copper conductors leading
to a higher release of copper into the fluid. A release
of copper into the fluid could have a negative impact
on the DDF value.
Fig. 9: Total acid number in natural ester and uninhibited mineral oil
From the DDF measurements on the aged fluids it

was observed that for the natural ester there was a
significant change in DDF value whereas the mineral
oil remained at low values even after 17 weeks. High
DDF values may deleteriously affect the power factor
and/or the insulation resistance of the electrical
equipment [14].
In the same context of evaluating the electrical
properties it was of high interest to analyze the
4.
CONCLUSIONS
The differences in oxidation stability requirements

for each fluid make a comparison complex. It can
generally be said that natural esters have weaker
oxidation stability versus mineral oil and synthetic
esters.
The high fire and flash point would be one of the
main benefits with the ester fluids, even though you
could question if 250C is enough if there is a severe
fault in the transformer.
To be able to make correct actions based on DGA
224

April 8-10
2015

results it is important to be aware of the different

gas formation patterns among the fluids. Today many
of the interpretation guides are based on mineral oil
but there are guides for non-mineral oil fluids as well
which should be used when appropriate. Care should
be taken when mixing different natural ester products
since this can affect the interpretation.
The conclusion of the study of the fluid/paper
insulating system when focusing on the properties
of the fluids are that ester fluids degraded much faster
when evaluating DDF and acid number and the release
of copper into the fluid seems to be higher; affecting
the dielectrical properties. It has earlier been shown
that the paper is preserved when ester is being used
as insulating fluid but it seems to be at the expense
of a very fast degradation of the ester.
BIBLIOGRAPHY
1.
CIGRE working group A2.35, (2010) CIGRE 436,

experiences in service with new insulating liquids
2.
American Society for Testing and Materials,

ASTM D6871-03 (2008), Standard
Specification for Natural (Vegetable Oil) Ester
Fluids Used in Electrical Apparatus.
3.
International Electrotechnical Commission, IEC

60296 ed 4 (2012), Fluids for electrotechnical
applications - Unused mineral insulating oils for
transformers and switchgear.
4.

61099 ed 2 (2010), Insulating liquids Specifications for unused synthetic organic esters
for electrical purposes.
5.

62770 (2013), Fluids for electrotechnical
applications- unused natural esters for
transformers and similar electrical equipment.
6.
A. Darwin, C. Perrier, P. Folliot, The use of natural

ester fluids in transformers, MatPost 07
7.
Duval M., The Duval Triangle for Load Tap

Changers, Non-Mineral Oils and Low
Temperature Faults in Transformers, IEEE
Electrical Insulation Magazine, Vol. 24 Issue 6
pp. 22-29, Nov.-Dec. 2008
8.
Hohlein, I. Kachler, A.J., Influence of moisture

and temperature on degree of polymerization
and formation of furanic compounds in freebreathing systems, Electrical Insulation
Magazine, IEEE 2005, 21 (5), 20-24
9.
Junru Xiang, Jian Li*, Zhaotao Zhang, Influence

of Water Content on the Aging Performance of
Natural Ester-paper Insulation, State Key
Laboratory of Power Transmission Equipment
& System Security and New Technology,
Chongqing University, Chongqing, China
10. Lars E. Lundgaard, Walter Hansen, Dag

Linhjell, and Terence J. Painter, Aging of OilImpregnated Paper in Power Transformers,. IEEE
Transactions on Power delivery 2008, 15, (2),
540-546
11. Lundgaard, L. E.; Hansen, W.; Ingebrigtsen,
S.,Aging of mineral oil impregnated cellulose by
acid catalysis, IEEE Transactions on Dielectrics
and Electrical Insulation 2004, 19, (2), 230-239
12. A. Kalantar*,t and M. Levin, Factors affecting
the dissolution of copper in transformer oils,
Lubrication Science 2008; 20: 223-240
13. Ruijin Liao, Shuaiwei Liang, Caixin SUN, Lijun
Yang, Huigang Sun, European transactions on
electrical power 2010; 20:518-533, A
compeditive study of thermal aging of
transformer insulation paper impregnated in
natural ester and in mineral oil
14. International Electrotechnical Commission, IEC
60422 ed 4 (2013), Mineral insulating oils in
electrical equipment - Supervision and
maintenance guidance
225

April 8-10
2015

Design of a Centralized Substation Synchronizing System

Dale Finney
Mangapathirao V. Mynam
Marcos Donolo
Schweitzer Engineering Laboratories, Inc.
ABSTRACT
Large substations often have complex and dynamic
topologies. The voltage available on either side
of an open breaker may originate from a number
of sources. This has led to the development of
centralized systems to carry out synchronismcheck functions to synchronize all breakers within
the substation. Such a system uses the status of
breakers and disconnects to identify a voltage
source for each side of the breaker that is to be
synchronized. Custom logic is required to
accommodate the topology of a particular
substation. In the past, these systems have been
realized using custom hardware or programmable
logic controllers (PLCs) and significant amounts
of wiring. This paper describes in detail a
synchrophasor-based approach that provides a
significant reduction in the effort and cost required
to design, build, and test a centralized
synchronizing system. Phasor measurement and
control units (PMCUs) transmit voltage phasors
and breaker and disconnect status to a central
controller. The central controller time-aligns the
data and selects the correct voltages to use for
synchronizing according to the present status of
the breakers and disconnects. Once the
appropriate checks of the voltages are made, a
close command is sent from the central controller
to the PMCU responsible for the breaker that is
to be closed. A primary objective is to reduce the
requirement for custom logic as much as possible.
The design relies heavily on using the program
organizational units (POUs) described in IEC
61131. These can be developed, tested, writeprotected by passwords, and easily reused in
subsequent projects.
Amy Sinclair
voltage magnitude and frequency in one island,

whereas synchronism check does not carry out
voltage or frequency control. The scheme
described in this paper is applicable to
synchronism check and can be extended to support
synchronizing two islands.
1.
INTRODUCTION
This paper describes the implementation of a

centralized synchronizing scheme for substations
with complex topologies. Substation configurations
exist that present challenges for synchronizing. One
such configuration is shown in Fig. 1. Assume that
Breaker 2 is to be closed. A synchronizing voltage
source for the top of Breaker 2 is provided by
Potential Transformer A (PT A) when Breaker 1 is
closed and by PT B if Disconnect 4 is closed. A similar
situation exists for sources on the bottom of Breaker
2. A distributed scheme can be implemented with a
dedicated synchronism-check device for each
breaker. In this case, each device needs to select the
voltage source from four sources based on breaker
The synchrophasor-based approach proposed in

this paper is also applicable when synchronizing
two power sources. This usually entails controlling
226
Fig. 1: Substation Showing Available Synchronizing Sources

April 8-10
2015

and disconnect status. The logic for the voltage

selection is unique for each breaker.
Complexity can be reduced by adding or relocating
PTs. For instance, in Fig. 1, if PT B and PT C were
moved to the left-hand sides of Disconnect 4 and
Disconnect 5, respectively, then the need for voltage
selection logic would be eliminated. However, it is
not always practical to place PTs in the optimum
location for synchronizing. For instance, placing a
PT on a gas-insulated switchgear (GIS) bus is likely
to be more expensive than placing it at the line
terminal air-to-gas bushing.
It is possible to implement a distributed scheme
wherein an intelligent electronic device (IED) is
dedicated to synchronizing each breaker. Each IED
would either have to measure voltages from all
required PTs (often not possible) or switch PTs
externally. The wiring and logic associated with each
distributed scheme when taken as a whole would
likely be more complex than a centralized scheme.
2.
CONVENTIONAL CENTRALIZED
SYNCHRONIZING SCHEME
Centralized schemes have been implemented to

address the issue described previously. Conventional
implementations are composed of two main modules,
as shown in Fig. 2.
could also be implemented using an off-the-shelf

programmable logic controller (PLC) and a
synchronism-check relay.
The advantage of a centralized scheme is that it
concentrates the hardware and associated logic
functionality in one location. A disadvantage is the
significant amount of wiring required to bring all of
the voltage and status signals to a central location.
In implementations where a supervisory control and
data acquisition (SCADA) control to close the breaker
is provided by a central remote terminal unit (RTU),
the synchronizing scheme could be located in the
same cabinet. Because the RTU usually requires the
same signals as the synchronizing scheme, this allows
signals to be shared.
Another disadvantage of the centralized scheme is
that a single point of failure impacts the
synchronizing capability of the entire substation.
3.
SYNCHROPHASOR-BASED CONTROL
Time-synchronized phasor measurements, also

known as synchrophasors [1], have been widely used
for visualization and postmortem applications such
as power system model validation [2] [3]. Phasor
measurement and control units (PMCUs) provide the
synchronized measurements. These timesynchronized measurements, complemented with
the advent of synchrophasor-based controllers
(SBCs), allow users to implement closed-loop
synchrophasor-based control schemes.
Closed-loop control schemes using synchrophasors
have been applied in the power system. Some of the
implementations in service today are the following:
z
Fig. 2: Block Diagram of Centralized Synchronizing Scheme
The logic module performs voltage selection.

Depending on the breaker to be closed, the logic
module selects two voltage measurements (Voltage
1 and Voltage 2) and passes them to the
synchronism-check module. The synchronism-check
module sends a SynchOK signal back to the logic
module if the conditions for synchronism check are
met. The logic module then routes a close permission
to the particular breaker to be closed.
In present implementations, both modules have been

implemented using custom hardware. The scheme
227
Islanding detection in distributed generation

(DG) applications uses phase angle
measurements at the DG location and the point
of common coupling and calculates the rate of
change of angle difference (slip) and the rate of
change of slip [4].
Remedial action scheme based on low-frequency
oscillations uses the power measurements from
two intertie transmission lines and measures the
low-frequency oscillations. The scheme sends
a command to disconnect the intertie
connection when the oscillations are associated
with negative damping [5].

April 8-10
2015

SBCs mainly provide the following functions: time

alignment (TA), built-in logic functions, and userprogrammable logic functions.
A. Time Alignment
Time alignment is a key function in the design of
synchrophasor data concentrators and controllers.
It allows for communications latencies between the
phasor measurement units (PMUs) and the controller
or data concentrator. The measurements are timetagged with a common time reference (typically
Global Positioning System [GPS]). The TA function
opens a time window (message wait time) where it
expects all the measurements with the same time
tag to arrive, independent of their location. Some
implementations force the device measurements that
arrive outside the message wait time to zero and flag
these measurements to represent bad quality. This

wait time is typically configurable and should be set
based on the communications latencies and
applications. For example, a smaller message wait
time is applicable for closed-loop control
applications, whereas for postmortem or data
archiving applications, a longer message wait time
may be acceptable.
B.
Built-in Logic Functions
Specifically for SBCs, the capability to perform

calculations or mathematical operations on the timealigned phasor measurements is critical. Additionally,
more advanced built-in functions are made available
in some controllers, and some of these functions
include the following:
z
Three-phase real and reactive power
Fig. 3: Processing Latencies
Fig. 4: Synchrophasor-based Centralized Synchronizing System
228

April 8-10
2015

Phase angle difference
Modal analysis
Substation state and topology processor (SSTP),

as described in Subsection B of Section IV
consists of two functions that are built-in features

(time alignment and an SSTP) and several program
organizational units (POUs) that have been
constructed using IEC 61131 programming
languages.
Processing the available logic functions at

deterministic low computation times is another key
requirement for SBCs, as it is for any controller.
Today, SBCs are available that can achieve
computation times in the order of 4 milliseconds.
The low processing time of the controllers allows
the implementation of closed-loop synchrophasorbased control schemes that require strict response
times (in the order of 100 milliseconds). Based on
the output of the control schemes, SBCs are capable
of sending a control command to the appropriate
device to take a control action. In some
implementations, the PMCUs provide synchrophasor
measurements and are capable of receiving the
control commands from the SBCs and taking
appropriate action. Figure 3 shows the potential
latencies that are involved in a synchrophasor-based
control scheme. Users must compare these latencies
with the timing requirements of the application for
any critical control scheme using synchrophasors.
C. User-Programmable Logic Functions

Some of the SBCs available today provide users with
IEC 61131-3-compatible support for multiple
programming languages [6]. These languages include
the following:
z
Structured text (ST)
Function block diagram (FBD)
Ladder diagram (LD)
Continuous function chart (CFC)
Instruction list (IL)
The standard provides the syntax and semantics for

the programming languages. Depending on user
familiarity with a particular language and the
complexity of the program, users can choose a
particular programming language.
4.
CENTRALIZED SYNCHRONIZING
SYSTEM USING SYNCHROPHASORS
The system shown in Fig. 4 consists of PMCUs, an

Ethernet network, and a central SBC. The SBC
PMCUs located in each bay are responsible for

measuring the voltage and frequency and
transmitting synchrophasors to the SBC. Note that
in Fig. 4, each voltage source has a dedicated PMCU.
Connecting multiple voltage sources to the same
PMCU requires a PMCU that can measure multiple
frequencies. PMCUs are also responsible for sending
breaker and disconnect status and receiving close
commands.
Using the time alignment of the data from multiple
PMCUs, the SSTP module constructs the topology
of the substation based on the user configuration
and the existing state of the breakers and
disconnects.
When a user chooses a particular breaker to be
closed, the voltage selection block selects incoming
and running voltages for synchronization using the
present topology. One key advantage of this system
is the capability to select the best available voltage
measurements based on the topology of the system
as determined by the SSTP to run the synchronizing
logic. The synchronism-check module checks that
the incoming and running sources are in phase and
that the magnitudes and frequencies of the two
sources are within limits (typically close to nominal).
The scheme then generates the close command,
which is routed to the breaker that is to be closed.
The SBC can receive synchrophasor data at up to 60
messages per second. The SBC can process logic at
rates of up to 240 Hz or four times per cycle at 60
Hz. In a synchronizing application, for a maximum
slip of 0.067 Hz, as specified by IEEE C50.12 and
IEEE C50.13 [7] [8], a 240 Hz processing rate equates
to a shift of 0.1 degree per logic scan. Thus, 100
milliseconds of latency represents 24 logic scans at
0.1 degrees of travel per scan, which is 2.4 degrees
of error in the actual angle difference. In applications
where synchronizing is carried out at a much higher
slip rate, the additional latencies introduced by this
scheme could hinder performance [9].
The example substation shown in Fig. 5 is used to
describe the new scheme. In this substation, PTs are
229

April 8-10
2015

Fig. 5: Example Substation
located only at the line terminals. Each line terminal

can be connected to either bus. Buses can be
sectionalized using disconnect switches.
The controller logic is subdivided into several
modules, which are described in the following
subsections.
A. Time Alignment
Time alignment is described in detail in Section III.
It is carried out automatically and ensures that all
downstream operations are made using timecoherent measurements.
B.
Substation State and Topology Processor
The SSTP gathers time-aligned synchrophasor data

along with the status of breakers and disconnect
switches from PMUs and PMCUs for substation state
and topology assessment. The SSTP uses these data
to identify measurement errors and improve
measurement accuracy.
(1) SSTP Structure
The SSTP module is organized into three main
processors (Fig. 6): the topology processor (TP); the
Fig. 6: SSTP Includes Topology, Current, and Voltage Processors

to Refine Measurements and Identify Measurement Errors
current processor (CP), which is not used for this

application; and the voltage processor (VP). The
topology processor processes breaker and disconnect
switch status to obtain the substation topology and
then makes this information available to the current
and voltage processors. The current and voltage
processors use the substation topology and the
synchrophasor data to detect measurement errors
and refine the current and voltage measurements in
real time.
Referring to Fig. 5, assume we are trying to
synchronize Node N4 to the rest of the substation
through Breaker B3. The incoming voltage is
measured by the PT at Node N4, and for the running
voltage, the SSTP takes the median of the voltages
measured at Nodes N6, N10, and N12. If one of the
voltage measurements is bad (e.g., the PT fuse has
failed), the median discards the bad measurement,
making the synchronizer more robust than when
using traditional methods. Note that averaging the
two good voltage measurements with a bad
measurement will not produce a quantity suitable
for synchronization.
The topology processor uses branch status
information to provide topology information for the
current processor (not used in this application) and
the voltage processor. The topology processor
determines the current topology and the voltage
topology by merging busbar nodes to create node
groups according to the closed status of the branches
in the busbar arrangement. To create the current
topology, the topology processor merges nodes when
the nonmetered branches are closed or when the
branch close status quality of the nonmetered branch
is false. To create the voltage topology, the topology
processor merges nodes when branches are closed.
230

April 8-10
2015

The current processor uses the current topology for

current measurement checks and refinement. The
voltage processor uses the voltage topology for
voltage measurement checks and refinement.
(2) Node Merging Process Example
Without the SSTP, custom logic would be required

to determine the voltages on either side of a breaker.
For example, the pseudo code in Fig. 7 presents the
logic required to determine the voltage at N3 in Fig.
5 (the running voltage needed to synchronize B3).
As stated previously, the topology processor uses

branch status to merge nodes. This allows node
voltages to be combined. To illustrate the node
merging process, consider Fig. 5. The bus
arrangement has 14 nodes numbered N1 to N14 and
17 branches numbered B1 to B17 in Fig. 5. There
are five metered branches (B3, B6, B9, B14, and B17).
Only Nodes N4, N6, N10, and N12 include voltage
measurements. The topology processor considers all
branches as merging branches to create the voltage
node groups. Table I shows the branch-to-node data
array for the voltage processor when all branches
are open. The array shows the From and To node
identification for each branch.
After a branch closes, the topology processor
replaces all instances of the To node ID with the From
node ID in the branch-to-node data array. For
example, Table II shows the new array after Branch
2 merges Node 2 and Node 3. In this case, the To
node ID is 3 and the From node ID is 2, as shown in
Table I. In Fig. 5, Branch 2 is connected from N2 to
N3. Thus, when Branch 2 closes, all entries that were
a 3 in Table I become a 2 in Table II (highlighted in
yellow).
When this topology is passed to the voltage processor,
it combines the voltage measurements available at
both Node 2 and Node 3. No other voltage
measurements are combined.
Fig. 7: Custom Logic Required for N3 if the SSTP Is Not Used
Unique logic would be required for each of the

breakers, and this logic would be specific to the
particular substation.
The SSTP logic can also be used as a front end to a
load-shedding logic application, where a particular
load can be automatically selected to be shed based
on the dynamically changing topology.
C. Arming Logic
Referring once again to Fig. 4, the arming logic
processes close requests. The logic, shown in Fig. 8,
is responsible for opening a window for synchronizing
and for rejecting close requests if synchronizing is
in progress on another breaker.
Table 1: Branch-to-Node Data Array for the Topology Processor when all Branches are Open
Table 2: Branch-to-Node Data Array for the Voltage Processor when all Branche 2 Merges Node 2
and Node 3
231

April 8-10
2015

Fig. 8: Arming Logic
Fig. 9: Voltage Selection Logic
232

April 8-10
2015

The arming logic resets after a fixed delay (30 seconds

in this example). In a practical implementation, this
logic may also be subject to site-specific
requirements regarding situations such as failed close
attempts or station-wide interlocks.
For our example substation, a PMCU is dedicated to
each breaker in Fig. 5. The output of the logic,
CB_TO_CLOSE, is a number ranging from 0 to 5 that
indicates the breaker to be closed.
In this example, close requests originate from the
PMCU associated with a particular breaker but could
also originate from another source such as an RTU
or local human-machine interface (HMI).
D. Voltage Selection Logic

The voltage selection logic chooses the incoming and
running voltages for the particular breaker to be
closed. This logic is shown in Fig. 9. The IEC 61131

ST programming language is chosen for this module
because it is more appropriate for this application.
Note that this logic is very simple because it is
receiving the node voltages from the SSTP. The nodes
on either side of a breaker are always the same. For
example, if Breaker B3 is to be closed
(CB_TO_CLOSE=1), then the node for the incoming
voltage (NI) is N4 and the node for the running
voltage (NR) is N3. N4 has a physically connected
voltage source. On the other hand, N3 does not have
a physically connected voltage source but derives
its voltage through the process of node merging.
E.
Synchronism-Check Logic
The synchronism-check logic shown in Fig. 10

receives the incoming and running voltages and
checks that the difference in angle, magnitude, and
Fig. 10: Synchronism-Check Logic
Fig. 11: Time-Advanced Closing Logic
233

April 8-10
2015

Fig. 12: Close Selection Logic
frequency is within limits. The first input to the AND

gate asserts when the absolute angle between the
incoming and running voltages is less than a pickup
threshold (10 degrees in this example). The remaining
three inputs to the AND gate assert when the
magnitudes and slip of the incoming and running
voltages are within a set band. If the limits are
satisfied, the logic gives permission (SynchOK) to
the close command logic. In Fig. 10, an additional
logic function (INBAND) has been developed to
further streamline the logic.
The closing logic is shown in Fig. 11. This logic
includes a feature to provide a time-advanced close
command based on angle difference and slip
frequency and the circuit breaker close time. First,
slip is calculated by subtracting the incoming and
running frequencies. The advance angle is equal to
slip 360 CBCT (circuit breaker close time in
seconds). This angle is compared with the actual
angle between the two sources (DeltaAng). The
comparison is less than or greater than, depending
on whether the slip is positive or negative.
Accounting for circuit breaker close time ensures that
angle difference is minimized at the instant the
breaker primary contacts close.
Although intended to address circuit breaker delays,

this feature could also be used to accommodate other
sources of delay.
F.
Breaker Selection Logic
The breaker selection logic routes the close command

to a particular breaker, as shown in Fig. 12.
As mentioned at the beginning of this paper, the
scheme is applicable both to synchronism check and
synchronizing. In a synchronizing scheme, similar
selection logic would also route raise/lower
commands (not shown) to particular generator
controls in the same manner. A PMCU would be
located at the generator controls in this case to
translate raise/lower commands to electrical contact
closures.
5.
ADVANTAGES
The proposed scheme has significant advantages over

the conventional scheme described in Section II. All
signals are exchanged over the substation local-area
network (LAN), so most hard-wired connections
disappear. Today, many modern protective relays
support synchrophasors and are wired to all of the
234

April 8-10
2015

voltage sources throughout the substation. These

same relays are also often wired with breaker and
switch status throughout the substation. Thus, in
substations where relays supporting synchrophasors
are applied, the need for additional hardware and
wiring is minimal. A redundant LAN architecture
coupled to redundant central controllers removes
single points of failure. The proposed scheme can
more easily be applied than a hard-wired scheme in
locations where voltage sources and breakers are
separated by long distances, such as the case of a
generator power house that connects to the grid
through a remote switchyard.
The proposed scheme can be designed for easy
modification and maintenance. The scheme logic can
be initially designed taking into consideration the
ultimate size of the substation. The logic intended
for future devices would initially be unassigned. If
the substation were subsequently extended to
include a new circuit and associated breakers, then
a PMCU could be added to include new voltage and
status signals. These new signals would be routed to
the SBC and assigned to unused logic inputs. Changes
would be required for the SSTP; however,
configuration of this module is more akin to setting
a protection function than to developing logic. Most
logic modules would require no modification. These
modules can be locked against editing after initial
design and testing.
The features described in this paper also make this
scheme easily adaptable to a different substation with
a different topology. Most of the effort is restricted
to configuration of the SSTP.
The SSTP also improves the quality of the voltage
measurements. When several voltage sources are
connected to a node, the SSTP calculates the node
voltage as the median of the available measurements.
6.
programming features. This results in simpler, more

modular code.
In a conventional application, considerable effort
would be required to develop the voltage selection
logic for each breaker. This logic would be unique
for each breaker. If the scheme was reapplied to a
different substation, then this logic would have to
be rewritten. This paper shows how the SSTP can be
used to replace custom logic. Configuration of the
SSTP amounts to defining nodes and branches
arguably a much simpler process with less potential
for error. This results in applications that are more
generic and thus more easily adaptable.
This effort represents a further step in the transition
to substation automation designs that leverage
advanced IEDs and communication to reduce cost
and complexity.
REFERENCES
1.
A. Guzmn, S. Samineni, and M. Bryson,

Protective Relay Synchrophasor Measurements
During Fault Conditions, proceedings of the
32nd Annual Western Protective Relay
Conference, Spokane, WA, October 2005.
2.
E. O. Schweitzer, III, D. Whitehead, A. Guzmn,

Y. Gong, and M. Donolo, Advanced Real-Time
Synchrophasor Applications, proceedings of the
35th Annual Western Protective Relay
3.
IEEE Power System Relaying Committee,

Working Group C-14, Use of Synchrophasor
Measurements in Protective Relaying
Applications, 2013. Available: http://www.pespsrc.org/.
4.
J. Mulhausen, J. Schaefer, M. Mynam, A.

Guzmn, and M. Donolo, Anti-Islanding Today,
Successful Islanding in the Future, proceedings
of the 63rd Annual Conference for Protective
Relay Engineers, College Station, TX, March
2010.
5.
J. V. Espinoza, A. Guzmn, F. Calero, M. Mynam,

and E. Palma, Wide-Area Measurement and
Control Scheme Maintains Central Americas
Power System Stability, proceedings of the
Power and Energy Automation Conference,
Spokane, WA, March 2013.
CONCLUSION
The scheme described in this paper performed as

expected during bench testing, showing that it is a
viable alternative to existing centralized approaches.
This paper demonstrates that the inherent timestamping provided by synchrophasor measurements
allows them to be effectively applied for critical
control functions in the power system.
The approach makes extensive use of IEC 61131
235

April 8-10
2015

6.
IEC 61131-1, Programmable Controllers Part

3: Programming Languages.
7.
IEEE Standard C50.12-2005, IEEE Standard for

Salient-Pole 50 Hz and 60 Hz Synchronous
Generators and Generator/Motors for Hydraulic
Turbine Applications Rated 5 MVA and Above.
8.
IEEE Standard C50.13-2005, IEEE Standard for

Cylindrical-Rotor 50 Hz and 60 Hz Synchronous
Generators Rated 10 MVA and Above.
9.
M. J. Thompson, Fundamentals and

Advancements in Generator Synchronizing
Systems, proceedings of the 38th Annual
Western Protective Relay Conference, Spokane,
WA, October 2011.

AUTHORS
Dale Finney received his BSEE from Lakehead
University and his MSEE from the University of
Toronto. He began his career with Ontario Hydro,
where he worked as a protection and control engineer.
Currently, Mr. Finney is employed as a senior power
engineer with Schweitzer Engineering Laboratories,
Inc. His areas of interest include generator protection,
line protection, and substation automation. Mr.
Finney holds several patents and has authored more
than 20 papers in the area of power system
protection. He is a member of the main committee
of the IEEE PSRC, a member of the rotating
machinery subcommittee, and a registered
professional engineer in the province of Ontario.
Mangapathirao (Venkat) Mynam received his MSEE

from the University of Idaho in 2003 and his BE in
electrical and electronics engineering from Andhra
University College of Engineering, India, in 2000.
He joined Schweitzer Engineering Laboratories, Inc.
(SEL) in 2003 as an associate protection engineer in
the engineering services division. He is presently
working as a senior research engineer in SEL research
and development. He was selected to participate in
the U.S. National Academy of Engineering (NAE)
15th Annual U.S. Frontiers of Engineering
Symposium. He is a senior member of IEEE.
Marcos Donolo received his BSEE from Universidad
Nacional de Ro Cuarto, Argentina, in 2000 and his
MSEE (2002), his master of mathematics degree
(2005), and his Ph.D. in electrical engineering (2006)
from the Virginia Polytechnic Institute and State
University. Since 2006, he has been with Schweitzer
Engineering Laboratories, Inc., where he is presently a
lead research engineer. He is a senior member of IEEE.
Amy Sinclair received her BSEE degree from Queens
University, Kingston, in 1989. She joined Ontario
Hydro in 1989, working for ten years as a protection
and control engineer in the areas of design,
operations, and project management. In 2000, she
joined ELECSAR Engineering as a project manager
with a focus on protective relaying and substation
design. Since December 2006, she has been
employed with Schweitzer Engineering Laboratories,
Inc. as a field application engineer located in
Chatham, Ontario. She has been registered as a
Professional Engineer of Ontario since 2001.
236

April 8-10
2015

Redundancy in Digital Substations

D.L.P. Jenkins
S. Richards
Alstom Grid
United Kingdom
S. Vigouroux
Alstom Grid
France
SUMMARY
The introduction of the process bus and station

bus to modern digital substations has raised many
questions and concerns. Protection Engineers are
responsible for ensuring that protection and
automation systems have an adequate level of
redundancy to be considered suitable for critical
power system applications.
Many new methods and technologies are now

available that offer the means of providing this
redundancy; however the temptation to introduce
many of these at once could lead to a design that
is unnecessarily complex, difficult to test and
unreliable in practice as a result.
For example it is now possible to design a process
bus architecture whereby protection relays
subscribe to multiple sampled value streams from
redundant merging units. However the question
of whether this is actually necessary is a valid
one. In a conventional substation such a design
philosophy would equate to protection relays
being wired to multiple redundant instrument
transformers, which is not common practice nor
indeed desirable.
This paper will review the contemporary
protection engineering principles that have been
used for many decades to engineer redundant
systems and then seek to establish how these same
techniques can be applied to new digital
substation architectures.
E. Avinash
Alstom T&D
India
The architecture of the process bus, station

bus and time synchronisation network. For
example how merging units are connected to
Intelligent Electronic Devices (IEDs) and how
connections are made between substation
bays to share sampled values and GOOSE
messages.
Where protection and automation functions
are distributed across the IEDs, merging units
and other devices in the substation.
The objective of this review is not to propose an

ideal or target architecture, since the design of
suitable solutions depends on the application in
question. Instead, it will seek to identify how these
methods need to be considered and what tradeoffs exist from their use.
Keywords: IEC 61850, Station Bus, Process Bus,
Merging Unit, IEC 62439, Time Synchronisation
1.
INTRODUCTION HOW HAVE WE

PREVIOUSLY ACHIEVED
REDUNDANCY?
The principle of redundancy is a long established

requirement for substation protection systems in
order to achieve reliable power systems.
Methods and technologies that will be reviewed

include:
A definition in the context of protection and

automation is the provision of sufficient duplicate
components so there is at least two independent
protection functions, each of which is sufficiently
capable of carrying out the required function on their
own. We often refer to these two systems as Main
1 and Main 2 (see Figure 1) [1].
Network redundancy protocols such as IEC

62439 Parallel Redundancy Protocol (PRP)
and High-availability Seamless Redundancy
(HSR).
This architecture is provided because 100% perfect

reliability of any single device is not possible to
achieve in the real world, therefore to improve
reliability of the complete systems to an acceptable
237

April 8-10
2015

Fig. 1: Fully Redundant Main 1 / Main 2 Line Protection Scheme [2]
level, we design so that the failure of a component

does not disable the overall protection system.
Although redundancy and backup are often confused
as the same thing, they are actually two distinctly
different properties; a redundant system will provide
a duplicate function that is of equal performance to
the primary, whereas a back-up system may still
provide protection but be of inferior performance.
One example of a redundant protection system
would be a Main 1 distance pilot scheme with a Main
2 stepped distance scheme. A back-up to Main 1
could be simple overcurrent protection, which does
not provide the same level of sensitivity or speed as
a redundant Main 2 system [2].
Some designs in a conventional protection system
that will achieve this redundancy would include:
Duplicated circuit breaker trip coils
Duplicated DC power supplies for IEDs and

tripping circuits
Separate current transformers, or duplicated

secondary cores
Minimising potential for common mode failure

between Main 1 and Main 2 systems, such as
maintaining physical isolation, and by using
different operating principles (this ensures that
the two systems are truly independent to one
another)
As will be discussed in this paper, we can often accept

a lower level of redundancy where back-up is
provided and we then assess that this will result in a
sufficiently reliable overall protection system.
Sometimes a lower level of redundancy would be
accepted where the consequence of failure is less
severe. For example, it is common for some
distribution systems to provide backup protection,
but not redundant Main 1 / Main 2 systems. Such an
assessment however is very subjective which is why
it is common for different design philosophies to be
used between different end users.
2.
DIGITAL SUBSTATION ARCHITECTURE

NEW CONSIDERATIONS
When we look at the architecture of a digital

substation, it is apparent that many new technologies
Fig. 2: Digital Substation Architecture
238

April 8-10
2015

are available. However, the overall principle of

protection redundancy must be maintained even with
this new architecture as it cannot be allowed that
the failure of a single device results in the complete
loss of system protection. For example IEDs can now
send commands to circuit breakers through the
process bus rather than with copper connections,
but the process bus should incorporate sufficient
redundancy so that the failure of a single Ethernet
switch does not result in the failure of a GOOSE trip
signal being transferred.
The architecture of the first example may in fact not

be any more reliable than the second when
considering a single device failure scenario, and it is
also probably more difficult to commission,
troubleshoot and maintain. In both cases there is no
single point of failure but the first example is
significantly more complex. This illustrates a key
point that although it is possible to implement more
sophisticated designs to add redundancy with new
digital substation technology, it does not necessarily
mean that it is always appropriate to do so [3].
As an example of how some Engineers have

approached this problem, consider the conceptual
design of the first example in Figure 3.
What then are the key new components to a digital

substation that will need to have redundant designs?
Referring back to the conceptual architecture in
Figure 2 we can see new points of common mode
failure that require consideration and will now be
analysed in detail:
Fig. 3: Options for Protection IED Subscriptions to Merging Units
Because any device can potentially subscribe to any

merging unit on the process bus, it is possible for a
design architecture like this first example whereby
the protection IEDs subscribe to multiple instrument
transformer sampled value streams via redundant
merging units. However the question of whether this
is actually necessary is a valid one. In a conventional
substation such a design philosophy would equate
to protection IEDs being wired to multiple redundant
instrument transformers, which is not common
practice nor indeed desirable, as it is accepted that
if one of the instrument transformer circuits were to
fail that the Main 2 system will provide sufficient
backup. Often the sampled values from the merging
unit do not need to be shared to other devices,
therefore a point-to-point connection could be
entirely suitable as shown in the second example of
Figure 3. This also aligns with the principle of the
two main protection systems being independent of
one another.
1.
The Ethernet network (both station bus and

process bus)
2.
The time synchronisation network
3.
The functional architecture of protection

functions, since these may now be distributed
across the substation within different IEDs
3.
SUBSTATION ETHERNET NETWORK

REDUNDANCY
Logically, we can group the two Ethernet networks

in a digital substation into two buses; the station
bus and the process bus. Though these are separated
as logically distinct systems they can be made of
different components with varying levels of
redundancy. For example the process bus
connections between some merging units and IEDs
could be point-to-point with no network redundancy
whereas others could make use of a redundant
protocol like PRP. There may even be components
that are used for both the process and station buses,
for example in the figure below GOOSE signals are
used in both process and station level signal
exchange from the protection IEDs.
The types of network redundancy technology that
can be used for the substation Ethernet network can
be grouped into three categories:
3.1 No Automatic Redundancy (point-topoint link)

In this case the Ethernet connection is a point to
239

April 8-10
2015

point link only. In the event of failure the connection

must be manually repaired to re-establish
communication. This could take the form of a single
Ethernet connection from an IED to a network
switch, or a single connection between merging unit
and an IED as in Figure 4.
Another example of passive redundancy is hot

standby, which is a term used for devices with a
back-up network interface that is only activated in
the case of the primary interface failing.
The advantage of these technologies is their
simplicity; they require very little configuration and
training to staff who may be new to networking and
digital substation principles. However they are not
a suitable choice where system down time is an issue,
for example an application that requires a consistent
sampled value stream, or a transmission application
where a few hundred milliseconds of protection
unavailability would pose an unacceptable threat to
grid stability.
3.3 Active Network Redundancy

Fig. 4: Point-to-Point Network Connection
Such architecture relies on back-up devices (Main

2) to provide redundancy to the system, rather than
redundancy of the Ethernet network. But this
approach could be perfectly acceptable since a failure
would have to occur on both the Main 1 and Main 2
simultaneously for the complete system to be
disabled.
3.2 Passive Network Redundancy

Some networking technologies offer redundancy but
will cause system outages for a certain period of time
while communication is re-established. For example
Rapid Spanning Tree Protocol (RSTP) which is very
commonly used in substation LANs, may take
hundreds of milliseconds to reconfigure when a link
is broken [4], during which time no signals such as
GOOSE can be exchanged between the isolated parts
of the network (Figure 5).
The international standard IEC 62439 defines two

network redundancy protocols that can be used for
substation Ethernet networks; PRP and HSR [5]. Some
other protocols do exist but are not interoperable
solutions and not considered in this paper.
These two technologies differ to other protocols in
that they are termed as bumpless whereby recovery
time in the case of failure is zero.
PRP (Parallel Redundancy Protocol) uses a doublestar architecture. Two messages are sent to two
different networks simultaneously. HSR (High
Availability Seamless Ring) as the name suggests uses
a ring architecture. Like PRP two messages are sent
from each device, but these traverse the same LAN
in opposite directions.
Compared to HSR, PRP has the advantages of:
z
It can support twice as many devices for the

same network bandwidth
It does not require that all devices in the system
Fig. 5: Spanning Tree Protocol
240

April 8-10
2015

Fig. 6: Parallel Redundancy Protocol
support PRP or HSR (devices can be singly

attached to only one network if full redundancy
is not required)
z
Flexibility of LAN structure since it is possible

for the two LANs to have different architectures
and use different technology
The disadvantage of PRP is that it requires more

investment in network components compared to HSR
since the network must be duplicated.
It is entirely possible and indeed recommended that
PRP and HSR networks are mixed with one another
depending on the application. For example LAN-A
shown in Figure 6 may actually be formed using a
HSR ring, or it may use another redundancy

technology such as RSTP. It is recommended to not
restrict the networking technologies that can be used,
so that Engineers may select the optimum solution
for the application.
The application requirement is the determinant factor
for what level of network redundancy is required. In
some systems there may be a high risk of Ethernet
network failure, or a separate Main 2 system may
not be provided, in which case a bumpless redundant
network protocol such as PRP might be required. In
other cases a simple point-to-point link could be
completely sufficient.
Fig. 7: High Availability Seamless Ring Protocol
241

April 8-10
2015
4.

TIME SYNCHRONISATION SIGNAL

REDUNDANCY
One new consideration with the introduction of the

process bus is the provision of accurate time
synchronisation signals. Process bus applications
require sampled value measurements to be
synchronised very accurately. If an IED receives
sampled values from different sources, for instance
voltage measurements from one merging unit and
current measurements from another, then protection
maloperation could occur if these samples are not
properly synchronised. Because of this risk, the same
redundant design principles also apply to the time
synchronisation network.
z
Redundant time sources
Currently the most popular means of providing a

source of global time synchronisation is using
satellite signals from the Global Positioning System
(GPS). This is a very accurate and cost effective
solution but has some disadvantages, namely it is a
single source of failure. In addition to the risk of
GPS system failure, there is the potential threat of
GPS signals being jammed or spoofed for malicious
purposes [6].
However not all applications require time signals to
be globally synchronised, so long as all samples are
synchronised to the same local clock. For example,
in applications that are limited to within the
substation such as bus bar protection, there is no
risk if communications to GPS satellites are lost as
the function can still continue to operate as normal.
In such scenarios redundant time sources may not

be necessary. This situation is even simpler in cases
such as feeder protection where all samples may
emanate from the same merging unit, so real time is
irrelevant to correct operation.
In the case where the samples being compared are
geographically dispersed, such as line current
differential, then global synchronisation is very
important. If the clocks at either end of the line are
not synchronised to one another then a differential
current may be wrongly observed resulting in maloperation, unless the relay scheme offers suitable
mitigation.
In such a situation it is necessary rely on back-up
protection functions in the case of time source failure
or to provide a redundant global time source. One
method is to use provide a diverse satellite
technology not linked to GPS. Currently the Russian
GLONASS system is the only commercially available
technology worldwide for such purpose, but in future
the European Galileo and other regional systems may
provide suitable alternatives. However common
mode failure is still possible, for example solar flares
can affect all satellite systems equally [7].
An accurate time source can also be provided
without a satellite based system by using a Caesium
atomic clock (Figure 8). This is a very expensive
solution but has been demonstrated as technically
feasible for situations that require extremely high
levels of redundancy, as the chance of common mode
failure is reduced by providing diversity in the
technology used.
Fig. 8: Redundant Time Source System Example
242

April 8-10
2015

Redundant time signal distribution
For the distribution of time synchronisation signals

there are two main methods; to use separate physical
cabling such as 1 Pulse Per Second (1PPS) signals, or
to make use of the existing station bus and process
bus Ethernet network infrastructure and transmit time
synchronisation signals via a packet based
technology such as IEEE 1588 PTP.
Packet based technology can offer a lower cost
solution if it can make use of the existing process
and station buses, but this can require that a great
deal of devices in the substation support the IEEE
1588 C37.238 Power Profile to maintain the required
level of accuracy of less than 1 microsecond,
including boundary clock devices such as bay
controllers. At present this can be difficult to achieve
but will improve as support for the Power Profile
grows.
Signal distribution via physical cabling methods such
as 1PPS have been in use many for many years and
are well proven. To minimise single points of failure
the time signalling for Main 1 protection devices
should ideally not be the same as for Main 2. If
economically justifiable this would entail the
construction of two separate time sync signal
distribution networks within the substation so that
the failure of one does not affect both systems
simultaneously. Where two time sources are
provided, this could be achieved with optical
multiplexers (Figure 8).
A completely redundant solution using physical
cabling can be very cost prohibitive and would
normally not be provided if back up protection
functions are available that do not depend on time
synchronised measured values.
5.
PROTECTION FUNCTION
REDUNDANCY
The location of the protection functions is not

restricted when using digital substation architectures,
as devices are not bound to the copper wires of
instrument transformers. Measurements can be
distributed freely across the substation and shared
between different devices and locations on the
process bus without additional wiring.
This means that we now have the possibility to
economically distribute additional redundant
protection functions to any location within the

substation that has access to the station and process
buses. For example:
1.
At the point of measurement acquisition within

merging units
2.
In passive standby IEDs that are activated in

the case of device failure
3.
In active standby IEDs that continuously

provide redundant backup to one or more
protection IEDs
5.1 Redundant Protection Functions at the

Point of Acquisition
Merging units in a digital substation take the function
of measurement acquisition away from protection
IEDs and put this function right up to point where
primary measurements are taken. There is the
possibility that these devices could also perform
additional protection and control tasks.
The argument is that this could increase overall
reliability, since if the communication link to the
protection IED or the device itself were to fail, the
merging unit could continue to operate
independently and provide some form of backup
protection (for instance an overcurrent protection
function as shown in Example 2 of Figure 9 below).
While this is a form of redundancy for specific
functions, it is not equivalent to a redundant Main 2
protection system since the performance it provides
is less than the performance from a complete
protection IED.
Whether this architecture would really provide
additional reliability in an outdoor substation where
merging units will typically be installed in the yard
is open for debate since the design of a device
suitable for such an environment, as opposed to the
substation control/relay room, can be compromised
if it must be sophisticated enough to perform
protection functions.
The relay room is a much more climate-controlled
environment; to control temperature and humidity
utilities typically take a conservative approach at
transmission levels, by installing heating in cold/
temperate countries, and air conditioning in hot
countries.
Devices deployed in outdoor marshalling kiosks in
243

April 8-10
2015

Fig. 9: Protection Functions Inside or Outside Merging Units
the yard do not have this luxury, in environments of

extreme cold occasionally the kiosks may have a
heater, but generally no heating or air conditioning
is provided. This means that yard devices such as
merging units need to withstand greater variations
and extremes in ambient temperatures, in addition
to making sure that moisture and or pollution in the
air produces no degradation. They may also subject
to transmitted vibrations when mounted in kiosks
attached to the switchgear assembly.
Because merging units should withstand these greater
environmental extremes than protection IEDs, they
need to be designed in such a way as to maximise
reliability and to secure the highest possible MTBF
(mean time between failures). This is best achieved
by good design principles: deploying FMEA (failure
mode effects analysis) and secondly by minimising
the component count.
Devices that must perform sophisticated protection
functions, in addition to measurement acquisition
and switchgear control, may need to carry additional
components, modules or processing. For example,
they may require an additional processor for logic
or ancillary tasks whereas a simple merging unit
could perform all its functions in an FPGA design
(Example 1 of Figure 9).
Given the recommended design for simplicity of
merging units and circuit breaker controllers, this
will tend to intentionally limit the additional

functions that those devices can perform. They are
designed to be specifically fit for purpose, as opposed
to devices which claim to perform many tasks to the
compromise of availability and long service life in
their primary function.
In summary, the motivation to keep yard devices
simple, and to perform ancillary or complex
functions in the traditional location indoor in the
bay, remains high. Therefore distributing redundant
protection functions to yard mounted merging units
may not necessarily deliver a more reliable system
in practice, and it could be for the system redundancy
to be enhanced that additional redundant IEDs
should be installed instead, as will now be discussed.
5.2 Passive Protection Function Redundancy:

Standby IEDs
An interesting possible new redundant protection
function architecture is to install an extra IED (MPx),
connected to the substation buses, which will be used
to replace the functions inside another IED in the
case of failure. The replacement of functions can
occur without having to physically touch any hardware.
In the example below, if a failure occurs in a Bay 1
protection IED (MP2). We could disable this device
and activate a spare that is already installed in the
substation.
244

April 8-10
2015

Fig. 10: Example of Standby IED
The following sequence would occur under this scenario:
Fig. 11: Standby IED Activation Process
1.
2.
During the installation phase a spare standby

device is installed in the substation that remains
inactive, but can one day be configured to
replace functions in one of several bays in the
case of failure. The device is connected to the
process bus, but does not have any
subscriptions enabled.
3.
The configuration of the faulted device is

retrieved from a central location which is
synchronised to the field equipment (to ensure
that the settings exactly match) and loaded into
the standby redundant IED.
4.
The device is placed into the mode defined by

IEC 61850-7-4 Ed2 Test Blocked. This allows
for test signals to be injected into the network
to prove that the configuration is correct.
GOOSE signals issued by the device will be
flagged as test so that subscribing switchgear
controllers know not to trip during this testing.
If a failure occurs (Bay 1) we first isolate this

device by disabling its process bus and station
bus interfaces. This can be achieved by turning
off the attached network interfaces in the
substation switches.
245

April 8-10
2015
5.

In this way the protection can be tested all the

way up to the switchgear control merging units
without having to operate primary circuit
breakers or by carrying out any secondary
injection.
protection functions automatically based on the

isolator position statuses.
The standby IED is taken out of Test-Blocked

mode and activated so that it now replaces the
protection functions that were disabled from the
initial device failure.
Taking these same principles a step further, the same

redundant IEDs could be configured to actively
subscribe to a number of different bays and provide
active back-up protection, rather than merely be
configured to provide protection in the event of
device failure.
This sequence could be performed remotely or

potentially even automated in future.
The standby IED allows for reduced downtime in
the case of device failure. Rather than waiting for
maintenance crew to travel to the substation and
replace the faulted IED the protection functions can
be restored very quickly and the faulted device can
be replaced at the next maintenance cycle. Thus
overall reliability of the complete protection scheme
can be improved.
The concept of passive protection function
redundancy is actually nothing new, and already
exists in some conventional substations where a
Main and Transfer Bus is provided, particularly in
North America. With this architecture (Figure 12) if
the protection relay in Bay 1 fails, the source for
this bay can be transferred from the Main Bus to the
Transfer Bus. The spare line relay then selects the
appropriate settings group to replace the faulted IEDs
5.3 Active Protection Functional Redundancy:

IEDs in Parallel
For example in Figure 13, the active standby relays

may incorporate several independent functions for
other bays in the substation. In the event of any
relay failure in any bay, protection would still be
provided. Such back-up could supplement Main 1
and Main 2 protection in each bay, or could be used
in cases where there is no Main 2 protection provided.
In future such principles may lead to more
centralised protection system architectures for
backup protection. The redundant device could in
the most extreme case, provide back-up for the
complete substation.
Testing such schemes, where functions no longer
reside within a single device will rely more heavily
on online simulation and isolation such as the new
features provided by IEC 61850 Edition 2. For
example, there are now functions that are distributed
Fig. 12: Passive redundancy in a conventional substation, spare line relay application
246

April 8-10
2015

across the substation that need to be isolated during testing, whereas previously isolation could be achieved
by simply turning off devices in a single bay. Clearly, careful work will be required across the industry to
make such architectures safe and practical.
Fig. 13: Parallel Redundant IEDs Example
6.
CONCLUSION DESIGNING THE

RIGHT LEVEL OF REDUNDANCY
Two independent process bus LANs. One for

Main 1 process information and the other for
Main 2. The buses are formed using an active
network redundancy protocol (HSR rings) that
can selectively share information between bays
such as circuit breaker position status for
interlocking.
Redundant PRP connections between the

protection IEDs and switchgear control merging
units (SC MUs). This shows how PRP and HSR
To summarise all the points that have been raised in

this paper with respect to redundancy for digital
substations, an example of one potential design is
now presented. From the figure we see:
z
Fully independent Main 1 and Main 2 protection

systems, each sufficiently capable of performing
independently.
Fig. 14: Example Redundant Digital Substation Architecture
247

April 8-10
2015

architectures can be mixed; each port is

connected to an independent HSR LAN so that
every IED actually publishes and subscribes
GOOSE to both a Main 1 and a Main 2
switchgear control merging unit. This is in line
with the philosophy of a typical hardwired
tripping scheme, where operation of any Main
protection should result in the operation of both
redundant circuit breaker trip coils.
z
To maintain simplicity the merging units (MU1

and MU2) publish sampled values to a single
LAN and are subscribed to by a single protection
IED. For example the MP1 protection relay only
subscribes to MU1 and not to MU2. Adequate
redundancy is instead provided by duplicating
the merging units and protection with
independent Main 1 and Main 2 systems.
All merging unit connections are made via a
switch so that measurements can be shared to
other bays. These connections are point to point
as it is deemed that additional network
redundancy is not necessary.
Two redundant time synchronisation systems
are provided, one with a source from GPS and
the other from GLONASS to provide some
diversity in technology. Each time source
system independently synchronises a single
Main protection system. In the unlikely event
of common mode satellite signal failure (both
GLONASS and GPS not being received) the
substation relies on backup protection that does
not depend on time synchronised values.
A bay of standby redundant IEDs is provided
that can be configured to quickly replace the
protection functions residing in any other Main
protection IEDs in the event of failure.
Therefore this example is but one potential solution

and is not the optimal for every single case; a tradeoff will always exists between cost and reliability.
In conclusion, redundant designs that add additional
complexity without meaningful increase in system
reliability should always be challenged and above
all, the established protection design principles used
in conventional substations relating to redundancy
are still valid for digital substation technology.
BIBLIOGRAPHY
1.
Network Protection and Automation Guide,

Alstom Grid, 2011.
2.
Redundancy Considerations for Protective

Relaying Systems, IEEE PSRC, WG19 Report,
2007.
3.
R. Hughes, Against and For Process Bus and

Sampled Values, Retrieved from: https://
ideology.atlassian.net, 2012.
4.
M. Pustylnik, M. Zafirovic-Vukotic, R. Moore.

Performance of the Rapid Spanning Tree
Protocol in Ring Network Topology,
RuggedCom, Inc
5.
IEC 62439-3, Industrial communication

networks - High availability automation
networks - Part 3: Parallel Redundancy Protocol
(PRP) and High-availability Seamless
Redundancy (HSR), Second Edition, July 2012.
6.
Volpe, J.A., Vulnerability Assessment of the

Transportation Infrastructure Relying on the
Global Positioning System, National
Transportation Systems Center, August 2001.
7.
Vladislav V. Demyanov, Yury V. Yasyukevich

and Shuanggen Jin. Effects of Solar Radio
Emission and Ionospheric Irregularities on
GPS/GLONASS Performance, September
2012.
It is important to observe that different levels of

redundancy are appropriate for different cases.
248

April 8-10
2015

POWERGRID Experience on Failure Prevention of

Transformers & Reactors
P.N. Dixit
R.K. Tyagi1
Gunjan Agrawal
Sumit S.H. Ray
Power Grid Corporation of India Ltd., India
ABSTRACT
Large power transformers are expensive and
critical elements of a utility transmission system.
Some transformers are so important that
consequences of their failure are tremendous e.g.
765kV Transformers and Converter Transformers
of HVDC System which are transmitting bulk
power from remote generating stations to load
centers. POWERGRID is operating and
maintaining about 650 Power Transformers, 90
converter transformers and 1000 shunt reactors
installed in 190 substations across the country.
Present task in front of POWERGRID is to
maintain serviceability of such a large fleet of
Transformers/ Reactors which are critical
considering the ageing fleet, large addition of
new Transformers/ Reactors especially at 765kV
level and significant effect on the grid due to any
outage.
During Operation and Maintenance of
Transformers & Reactors in POWERGRID, various
problems/ failures are observed. These failures
were analyzed in detail and various measures are
taken to improve the serviceability of
Transformers & Reactors. In this paper, few case
studies are brought out where failure prevention
of Transformers & Reactors are achieved based
on past failures/ problems encountered in
POWERGRID.
1.
INTRODUCTION
Transformer is a completely sealed unit and limited

O&M activities can be performed in such units once
it is installed. All actions are to be taken by
1
tyagir@powergridindia.com
manufacturer during design, manufacturing,

assembly, drying and testing stage so that the
Transformer provides satisfactory service. In case,
weak areas are left, the Transformer will have reduced
life and will give continuous problems. Generally, it
is said that the life of the Transformer is the life of
paper insulation and change of its condition with
time under impact of thermal, electrical (die-electric),
chemical, electromagnetic and electro dynamic
stresses as well as impact of ageing processes.
However, there are several other aspects which needs
to be considered especially in the early years of
commissioning and can affect the life of the
Transformers.
Major critical areas identified during Operation of
Power Transformers and Shunt Reactors in early years
of commissioning in POWERGRID network are:
High failure rate of Bushings leading to fire

incidences.
High moisture content in cellulosic insulation

left in the transformer insulation due to
inadequate dry out either at manufacturers work
or at site during commissioning resulting in fast
deterioration of oil parameters and accelerated
ageing of cellulose
Manufacturing issues reflected in fault gases/

leading to failure
Transportation issues reflected in core isolation

and gassing
Switching of 765/400kV Transformers &

Reactors at wrong instances leading to humming
sound in transformers
POWERGRID has taken various measures to improve

the performance of Transformers and Reactors based
on above operational feedback. In this regard, various
249

April 8-10
2015

modifications were carried out in the Technical

Specification and maintenance guidelines were also
reviewed and changed. The failures investigations
were carried out in association with manufacturers
and based on investigation, various improvements
were carried out in design, manufacturing,
workmanship, testing, transportation, erection and
commissioning of the Transformers.
2.
FAILURE PREVENTION OF BUSHINGS
Bushings are the most critical component in Power

Transformers as the failure of Bushings often leads
to fire and subsequent major failure of the
Transformers and Reactors. In most of utilities,
Bushings are main cause of Transformer failures. As
per failure statistics of Transformer and Reactor in
POWERGRID, Bushings are responsible for about 41
% of the total failures of Transformer and Reactor.
Failure of EHV bushing is often followed with
catastrophic consequences as explosion, tank
rupture, fire, etc. During detailed analysis of the
failures, the reasons of failure identified were
inadequate care taken during manufacturing,
inadequate design, problem in leads, test tap design
problem, improper erection etc.
Some of the case studies of Bushing failures along
with measures taken to prevent the failures are given
below:
2.1 Problem of Flashover in 800 kV OIP

Bushings
A particular model 800kV bushing exhibited external
flashover between top head container i.e expansion
chamber (Fig. 1) to bushing flange (Fig. 2) during
drizzly rain condition. Flashovers in the bushings
were encountered when installed at an angle of 15
Fig. 1: Flashover on expansion chamber
degree from vertical with deposition of dust layer

on porcelain surface. Initiation points of flashover
were found on the container below the welded joints
of the RIV shield. Five incidences have been
encountered so far at 2 numbers of 765kV S/s where
the initiation point was exactly same with inclined
installation of that particular bushing.
Initially, it was suspected that flashover was due to
arrangement of RIV shield (especially the upper ring)
as shown in Fig. 3 causing collection of more
rainwater and depositing on the upper surface of the
container. This water then tends to flow in the areas
where the flashover points occurred.
Rain pollution test and RIV test simulating site
conditions were carried out at Manufacturer
works. Rain pollution test was carried out by pre
stressing at 1.15 Um (518kV) for 10 minutes
followed by test voltage up to 530kV for 20 minutes
with simulating the site condition with following
combination:
(i)
On the original bushing with original RIV shield.
(ii) On the original bushing with modified RIV

shield.
(iii) On the modified bushing having reduced
diameter of porcelain with original RIV shield.
(iv) RTV coated original bushing with original RIV
shield.
(v) On the original bushing with polymer housing.
Significant levels of external discharge activity were
observed within one minute after the beginning of
rainfall application in case (i), (ii) on the lower part
of the porcelain. In case of (iii), (iv) & (v), there was
no external discharge activity during application of
test voltage under polluted and wet conditions.
Fig. 2: Flashover on flange
250
Fig. 3: RIV arrangement

April 8-10
2015

The external flashovers were caused by buildup of

pollution on the surface of the porcelain, which
become wet when exposed to rainfall. This formed
conductive layers on the surface of the porcelain,
which discharged externally in the presence of high
voltage electrical field in the service condition.
Because the bushings were installed at an angle, the
length of the polluted surface was extended
compared to bushings installed vertically.
Additionally, as the rainwater fell onto the surface
of the porcelain it was able to move further along
the hidden upper surface of the porcelain sheds on
inclined bushings, due to the decreased inclination
of this surface from the horizontal. This further
served to increase the size of the conductive bands
and hence reduce the insulation performance along
the surface of the conductor.
temperature. Surprisingly, the Tan delta of the

problematic Bushing was at par with the tan delta of
the other healthy Bushings. It was then noticed that
the Tan delta of the problematic Bushing was varying
with the temperature with very high values at high
temperature and normal reading at ambient
temperature. Other two bushings however were
showing nearly stable Tan delta values at different
temperatures. Then, the measurement of Tan delta
was started just after taking the outage of the unit
and thereafter interval of every three hours. As shown
in Fig 4, R phase bushing is showing large variation
in Tan w.r.t. temperature. However, other two
bushings are showing stable results.
For immediate action , application of RTV coating

were made on these bushings where pollution level is
more to creates a hydrophobic layer on the surface of
the porcelain, which effectively prevents the formation
of conductive bands on the insulator surface even
during wet polluted conditions.
2.2 Bushing Diagnosis by Capacitance and

Tan Delta Measurement at Variable
Temperature
One 50 MVAR Reactor was in operation for three
years. Rise in fault gases (esp H2, CH4, CO & CO2)
observed after three years of service. DGA results
are shown in Table 1. During normal maintenance
program, Bushing tan delta of that Reactor was
measured immediately after shutdown (i.e. at higher
temperature). It was found that one bushing was
showing very high Tan delta. The confirmatory test
was carried out after some time i.e. at normal ambient
Fig. 4
The oil sample of the problematic bushing was also

taken and very high fault gases observed in the
bushing. The defective bushing was replaced with
healthy one and after replacement; DGA trend of
shunt reactor was found stable and presently running
without any problem. DGA results of R phase
bushing are shown in Table 1.
Table 1: DGA Results of Shunt Reactor main tank oil & Bushing
Sample Date
4.4.2008
19.04.2008
4.5.2008
3.11.2008
12.2.2009
6.10.2009
H2
0
28
90
285
350
558
CH4
0
2
2
12
21
56
C2H4
0
1
1
3
3
3
C2H6
0
0
0
2
3
6
C2H2
0
0
0
0
0
0
CO
7
50
74
205
223
351
CO2
155
1821
3496
6770
7474
12708
3.2.2010
07.08.2012
Bushing DGA
11
33
18614
3
9
1750
0
3
1
0
2
114
0
0
0.57
16
282
158
684
6212
593
251

April 8-10
2015
3.

PROPER VACUUM DRYING PROCESS

TO AVOID FAULT GAS GENERATION
In certain cases it has been observed that fault gasses

especially H2 and CH4 increases abnormally within
a short span of commissioning or re-commissioning
and subsequently stable after few months. DGA of
the few such units were analyzed through different
ratio analysis as stated below and it was observed
that the H2 and CH4 generated due to PD inside the tank.
inrush currents were observed along with high level

of humming noise even after controlled energizing
at voltage peak. The available system disturbance
recorder was triggered to record with manual closing
of the circuit breaker and switching phenomena was
recorded. It was confirmed that the implemented
configuration in the controlled switching device
needs review. In the implemented case, the device
had been targeting individual voltage peak points of
Unit-I
Unit-II
June14
March15
Nov13
H2
1614
18
2857
CH4
70
6
184
C2H4
1
1
6
C2H6
5
1
23
C2H2
0
0
0
CO
208
176
434
CO2
280
604
1696
Fuzzy Interpretation
PD
Normal
PD
Rojers Technique
PD-Corona
PD-Corona
Duval Triangle
PD
T1 fault
IEEE Rojers Ratio No matching
PD
In all the above cases, Partial discharge test and
internal inspection couldnt reveal any abnormality.
These units were taken into service after carrying
out proper vacuum drying and adequate settling time
for oil and no further rise in fault gases appeared
even after one year of re-commissioning.
POWERGRID experienced such type of problem in
number of cases where H2 & CH4 gas disappears
after proper vacuum drying and with adequate
settling time for oil during re-commissioning.
From the above it may be envisaged that any
abnormal rise in H2 & CH4 in a short span of
commissioning may be because of Partial Discharge
due to presence of trapped air. It may be due to
improper vacuum drying process and settling time
during new commissioning/ after any repair work.
In such cases, adequate vacuum drying out and
settling time may resolve the problem.
4.
USE OF CONTROLLED SWITCHING

DURING ENERGIZATION OF 765 / 400 /
33 KV, 1500 MVA TRANSFORMER
BANK
During energisation of 765/400/33kV Transformer

bank in one of a Substation in POWERGRID, High
Dec14
45
15
3
3
0
221
1191
Normal
Unit-III
Dec12
1655
127
3
18
0
1104
3360
PD
PD
T1 fault
PD
March15
85
64
4
11.4
0
765
2865
Normal
the phases for closing instant of the respective circuit

breaker poles. This is an ideal case for a Y-Y
transformer. The 765 / 400 / 33 kV, 1500 MVA
transformer banks in question comprised of single
phase two winding units with main winding in star
formation with other units of the bank and another
winding (tertiary) in delta formation with other units
of the bank. The controlled switching device
required to be configured taking into account the
effect of common delta tertiary winding formation
for the transformer bank.
For the implemented controlled switching device
configuration, the device had been targeting
individual voltage peak points of the phases for
closing instant of the respective circuit breaker poles
with taking into account equal pre arcing times for
all three circuit breaker poles. However, for the
present case, with the closing of first pole of the
circuit breaker, the common delta tertiary winding
of the transformer bank shall get energized and in
turn there shall be transformation voltages on the
main windings other units of the transformer bank.
Hence, for the second and third poles of the circuit
breaker (closing later), there shall be a slightly
different pre-arcing time than that of the first circuit
252

April 8-10
2015

breaker pole. As the transformed voltages appearing

on the main windings of the other two units of the
transformer bank (appearing on the load side of the
circuit breaker interrupters) shall be in phase with
the source side voltages (which shall extend after
closing of the circuit breaker), the pre arcing times
for second and third circuit breaker interrupters were
deemed to be minimal.
Hence, it was decided to alter the pre acing times
for the second and third circuit breaker poles which
the controlled switching device takes into account
for switching (closing of circuit breaker) at the desired
instant. The pre arcing times set for second and third
CB poles were reduced from 2.5 m sec to 0.4 m sec.
Controlled closing of the circuit breaker was
performed with new configuration of the controlled
switching device and minimum inrush currents were
recorded during energization of the transformer
banks. Humming noise levels which were high during
earlier transformer banks energization also subsided
to normal level.
DR for energization of 765 / 400 / 33 kV 1500 MVA
transformer bank with prior settings of controlled
switching device; phase currents above 1000 A rms
values are seen
DR for energization of 765 / 400 / 33 kV 1500 MVA

transformer bank with revised settings of controlled
switching device; phase currents 10-12 A rms values
given below
Thus, with the close analysis of Disturbance recorder,

the problem was identified and corrective action was
taken. Earlier, provision of Controlled Switching
devices was only on the 765kV side of Circuit
Breakers. It was however observed that in case of
newly commissioned 765/400 kV Substations,
Transformers have to be energized some of the
instances from 400kV side also. Hence, now the
253

April 8-10
2015

Controlled Switching devices are provided on 400kV

side Circuit Breakers also.
5.
CONCLUSION
Proper Condition Monitoring of Transformers &

Reactors is very essential for timely detection of
problems and necessary corrective action. In
POWERGRID, with the help of state of the art
Condition Monitoring techniques, large numbers of
Transformers & Reactors have been saved from
possible failures. This has resulted in significant
reduction in failure rate of Transformers & Reactors.

It is also observed that some of the problems are not
getting detected by one or two tests and additional
tests may be required to find out the exact problem.
Acknowledgement
The authors are grateful to POWERGRID
management for granting their permission to write
and present this paper. The views expressed in this
paper are of the authors only and not of the
management of POWERGRID.
254

April 8-10
2015

Advancement and Broadening of Controlled

Switching Technology
M. Stanek1
ABB High Voltage Products
Switzerland
U. Parikh
ABB High Voltage Products
India
SUMMARY
Controlled switching, also known as point-onwave switching or synchronous switching, is a
well-established technique for minimizing
switching transients during closing and opening
of a high-voltage circuit breaker. The latest
generation of point-on-wave (POW) controllers,
designed on modern hardware and software
architecture, can offer additional benefits to the
user. These include,
z
Predefined (built-in) and user defined

switching strategies.
Enhanced current and load voltage based

switching detection for high-accuracy
monitoring and adaptive correction for any
load configuration.
Condition monitoring of circuit breaker

including electrical and mechanical ageing,
without the need for additional sensors, to
assist in asset management.
Simple wizard stylePC tool, including a

library of circuit breaker models, to assist the
user in configuring the device.
IEC 61850-8-1 with redundancy andbuilt-in

cyber security.
Voltage and current signals acquisition from

IEC 61850-9-2 (LE) sampled value inputs and/
or conventional analog inputs.
Monitoring and diagnostic support through

graphical local and remote user
interfacesincluding web browser.
michael.stanek@ch.abb.com
A. Parapurath
A. Talluri
ABB Substation Automation Products
India
These state-of-the-art features of the POW

controller will increase its value to the user at no
extra effort.
Keywords: Controlled switching, point-onwave controller, circuit breaker, monitoring, user
interface, IEC 61850, digital communication, cyber
security.
1.
INTRODUCTION
Controlled switching is employed to reduce the

stresses on the circuit breakers as well as on the
system and equipment to be switched during
energization and de-energization operations. During
energization of capacitor banks and transformers,
thermal stresses are imposed due to high amount of
inrush currents and also result intemporary
overvoltages. During de-energization of reactors and
capacitor banks the dielectric stresses are imposed
on circuit breakers due to transient recovery voltage
appearing post successful de-energization.
Energization of transmission lines imposesswitching
overvoltages. Controlled switching (point-on-wave
switching, POW) can reduce these stresses by
energizing and de-energizing the loads on specific
points on the reference voltage or current waves.
The targets shall be selected based upon the
connection and design configuration of the
equipment to be switched, which indirectly takes
care of electrical and magnetic coupling between the
phases as well. The targets shall also be selected in
accordance with the chosen circuit breaker, taking
into account its performance in terms of mechanical
and dielectric characteristics, including any
influences from external parameters on operating
times.
255

April 8-10
2015
2.

EMERGING TRENDS IN CONTROLLED

SWITCHING TECHNOLOGY
Employing modern numerical relay technology in

POW controllers, it is possible to go for higher
performance in computations, which can assist in
achieving desired target accuracies and fast response
times. Furthermore, it is possible to choose desired
sensitivity of reference & feedback signals, in
particular where secondary measurement quantities
are quite insignificant. For some applications it is
required to detect a change in voltage or current signal
properties to estimate the makingor interruption
instants of a circuit breaker to assist operating time
adaptation.
As an example, Figure 1 shows energization of
anYNdconnected transformer from the high voltage
side. Phase L1 is energizedat instant 1; at that time,
Fig. 1: Voltage step change during controlled energization of

aYNd transformer bank.
voltage is induced in the other phases. Phases L2

and L3 are energized simultaneously at instant 2,
as seen from the current traces.Checking just the
presence of transformer voltage would not allow
correct detection of the L2 and L3 energization
instants. Rather, this requires identifying the knee in
each voltage trace using an advanced signal analysis
algorithm.
The disturbance recording facility available in
numerical relays makes it possible to analyze the
results and can also provide directions to improve
the performance further, provided the sampling rate
is high enough e.g. 80 samples per cycle. The effects
of external noise, inter-phase interference caused by
switching operations, etc. can also be eliminated
using advanced algorithms.
Furthermore, achieving the desired accuracy highly
depends upon the consistency of the circuit breakers
dielectric and mechanical characteristics. The scatter
in operating times of a circuit breaker and variation
in rate of rise of dielectric strength (RRDS) and rate
of decay of dielectric strength (RDDS) can highly
impact the accuracy in achieving the desired targets.
The effect of above mentioned parameter scatters
during a closing operation can be seen from Figure 2
below. See also [1] for further information.
The speed of the breaker during opening and closing
operations would also decide whether certain targets
can be achieved or not. For example, with slow
closing speed of a breaker, targets very close to gap
voltage zero cannot be achieved. Also, the effect of
Fig. 2: Scatter of various parameters during circuit breaker closing
256

April 8-10
2015

external parameters like DC voltage, gas pressure,

temperature, idle time, etc. shall be compensated
while computing the target if necessary to obtain
desired target accuracy.
In context to adaptation for mechanical and electrical
operating time variations, appropriate feedback
signals like load currents and/or voltages and
changeover of auxiliary contacts of the circuit
breaker can be arranged. The behavior of the
feedback signals shall also be considered while
choosing the type of feedback signal for adaptation.
As an example, for closing application of a grounded
reactor, all above mentioned feedback signals can
be used for adaptation. Whereas for transformer
energization application, the switching is to be done
on no load, for which the current will be too small
and will contain multiple harmonics. This will make
it unsuitable as adaptation feedback for operating
time variations. Hence, the configuration of the relay
in context to operating time adaptation will be
different and challenging for various type of
applications. In such circumstances, automated
choice of adaptation feedback based on system
configuration can be achieved with latest generation
of numerical relays.
Furthermore, the accuracy of measuring equipment

(CT & PT) and consistency of auxiliary contacts to
replicate the actual operating times plays a vital role
to avoid mal-operation of relay due to inaccurate
feedback signals. By mathematical smoothing of
computed errors, this effect can be taken care.
Latest generation of controllers having advanced
computation capability can create signature patterns
for various operating parameters of the circuit
breaker based on data collected from a number of
operations. These patterns can be stored and also be
used as a base to analyze and compare the
performance of the controlled switching for future
operations.
3.
CIRCUIT BREAKER MONITORING
Most signalsnecessary for controlled switching can

be re-used for condition monitoring and supervision
of the circuit breakers switching performance and
aging. Various monitoring parameters in context to
controlled switching performance are errors in
electrical andmechanical targets, electrical make and
break times,or mechanical closing and opening times,
as shown in Figure 3 below.
Fig. 3: Controlled energization performance and errors
257

April 8-10
2015

The introduction of high-end processors and new

algorithmsallows comprehensivemonitoring of
controlled switching operations for accuracy (target
errors, arcing time, etc.) and CB performance
(operating times, velocity etc.).This information is
used to optimize the controlled switching
performance, to detect undesired conditions such
as re-ignitions, and to aid in taking corrective actions.
If auxiliary contacts are connected to precision
binary inputs, mechanical CB performance can be
monitored very accurately as well. These monitoring
functions will also assist with asset management by
tracking electrical ageing(interrupter wear) and
mechanical ageing(number of Close-Open cycles
performed), so that necessary maintenance can be
scheduled well in advance.
4.
CIRCUIT BREAKER OPERATING TIMES
For successful controlled switching, the POW

controller needs to know the expected electrical
switching times of each CB pole. Electrical making
times are calculated from the mechanical closing
times and the expected pre-arcing times. Similarly,
the electrical interrupting times are computed from
the mechanical opening times and the expected
arcing time (usually until the next natural current
zero).
Ideally, a CB used for controlled switching will exhibit
low fluctuations (scatter) of its electrical and
mechanical switching characteristics. RDDS/RRDS
and compensation curves can be specified in advance
for a given CB model; from these data the controller
can calculate the target pre-arcing and arcing times
directly. The absolute value of closing and opening
time, however, will be different for each CB pole
due to tolerances in manufacturing and/or
adjustment. Traditionally, these individual values
have been measured on site during CB
commissioning, using dedicated equipment, and
manually programmed into the POW controller.
More recent concepts provide for the POW controller
to measure the mechanical operating times of the
CB by itself. This requires nothing more than
temporary electrical connections to the primary
contacts, for acquiring the instants of contact touch
or separation.
In the direct method, a DC voltage source is hooked
up to each interrupter and the signals are fed into
precision binary inputs of the controller. The

controller is set to issue appropriately timed
switching commands to the CB preferably staggered
by some hundred milliseconds, to allow distinct
allocation of signal status changes to each CB phase
and measure the changeover instants of the
contacts. If auxiliary contacts are connected for
monitoring purposes, as explained above, the actual
timing of those is recorded at the same time. With
suitable logic, this technique even allows detection
of wiring errors in the control circuit, e.g. wires
swapped between phases, between NO and NC
auxiliary contacts, or between Close and Trip circuits.
After every shot, the results are checked for
plausibility by the controller itself and presented to
the operator for further assessment. See Figure 4 for
an example. Minimum five acceptable Close and
Open operation cycles should be performed to assure
statistical significance. After completion of the test
series, the results are averaged to yield the expected
mechanical operating times of each CB pole.
Fig. 4: Example presentation of CB closing time measurement on

local display of POW controller.
5.
APPLICATION SETUP
A POW controller is usually delivered with default

settings, unless agreed otherwise with the supplier.
Prior to first use, it must be set up to match the
specific application:
z
258
Power system parameters (nominal system

voltage, frequency, grounding)
Circuit breaker properties (mechanical and
electrical
switching
characteristics,

April 8-10
2015

compensation curves, )
z
Load type and configuration (e.g. power

transformer, 3-limb core, YNd vector group)
Breaker control, i.e. which operations (Close/

Open) to control, and contingency in case
controlled switching is not possible
Configuration and ratings of reference signals

(from VTs, CTs, or even non-conventional
instrument transformers NCITs / Merging Units)
Number and characteristics of sensors

connected for compensation
What alarms or warnings to raise on what

conditions etc.
Traditionally, the user had to manually enter the

respective settings, based on the equipments user
manual, which was a tedious and error-prone
procedure. In the worst case, the data could not even
be saved as backup, or re-used for a similar
application, but had to be entered anew every time.
A modern POW controller will guide the user through
the process of entering all required information,
offering explanations and performing validity checks
on every value entered. It should not require to
assemble the application configuration in different
tools, but a single dedicated tool should do the job.
Where possible, only high-level information should
be asked. For example, the average user will be much
more comfortable specifying the type and
configuration of the load than having to manually
determine and enter the target energization angles

for the given load type in each phase. The latter task
is preferably left to an appropriate algorithm, of
course with the option to review and manually
correct the results if necessary.
An example screenshot of a wizard-style
configuration tool is shown in Figure 5.
6.
USER INTERFACES
Ideally, controlled switching is transparent to the

user. Once it has been put into service, the POW
controller should perform its duty without human
interaction, continuously supervising its own
performance and that of the controlled circuit
breaker. In case of an irregular condition, it will alert
the operator and provide all information necessary
for diagnostics.
In addition to alarm contacts, a state-of-the-art POW
controller will provide two kinds of Human/Machine
Interfaces (HMI):
1.
Local HMI (LHMI) on the device itself. As a

minimum requirement, the LHMI will comprise
several LEDs, to indicate the status and
condition of the controller and the CB at a
glance. In addition, many plant operators will
appreciate an interactive LHMI, comprising a
graphic-enabled screen and appropriate user
controls such as pushbuttons. Through this
LHMI, the user can quickly obtain detailed
information on recent controlled switching
operations and other events, without the need
to hook up a computer.
2.
Communication interface(s) for local or remote

access. This is important for several reasons:
(a) Access to information without the need to
be physically present at the device. This is
particularly relevant in unmanned
substations.
(b) Storage and backup of information, which
otherwise may be lost.
(c) Presentation of (particularly graphical)
information on a computer monitor is far
more user friendly than on the small screen
that would fit on the front panel of a
controller device.
Fig. 5: Example of wizard-style configuration tool for POW

controller.
(d) On a computer in the control center or even
259

April 8-10
2015

in the office, information can be compared

between devices, aggregated (e.g.
averaging, trend analysis), and visualized
as desired.
Ethernet-based communication interfaces open up
further possibilities:
z
IEC 61850 enabled devices can be integrated

with a digital station control system with
minimal effort and wiring.
A web HMI allows access from any location that

is connected to the station network (protected
by appropriate security measures) and from
different kinds of devices (e.g. PC, tablet),
making the user independent of product specific
software. An example of a graphical web HMI
is shown in Figure 6.
IEC 61850 features that allow integration of a POW

controllerin a digital substation automation system
include,
GOOSE Horizontal communication to exchange
information with other IEDs e.g. close and open
commands from a bay controller; transmit status
information and alarm signals; analog values for
compensation from an IEC 61850-8-1 enabled
remote I/O device.
Process bus for sampled values Enables receiving
current and voltage signals in digital format from an
IEC 61850-9-2 (LE) compliant merging unit.
MMS Vertical communication with SCADA for
control and monitoring.
Time synchronization Station level synchronization
(e.g. SNTP) for event time stamping; hardware
synchronization (1PPS) for handling process bus
data.
7.2 Communication Redundancy
Fig. 6: Example of web HMI showing waveforms of a controlled

closing operation.
7.
Reference voltage and current signals can be

received from digital merging units via the IEC
61850-9-2 (LE) protocol. This enables the use
of electronic voltage or current sensors for
controlled switching without further signal
conditioning.
DIGITAL COMMUNICATION
7.1 Digital Substation using IEC 61850

The general benefits of a digital substation
automation system using IEC 61850 are explained
in [2].
When information is exchanged through a single

communication channel, damage to the cable or
associated connections can force a shutdown of the
substation. So for critical infrastructure it is important
to have redundant communication channels: If
information is exchanged simultaneously through
two separate channels, the substation automation
system can continue functioning during a service
interruption in one of the channels. IEC 61850
recommends two ways for achieving communication
redundancy. They are PRP (Parallel Redundant
Protocol) and HSR (High availability Seamless Ring)
[3]
. Modern point on wave controllers enabled with
PRP will satisfy the communication redundancy
requirements in substation automation system.
7.3 Cyber Security

Introduction of Ethernet (TCP/IP) based
communication protocols, resulted in possibility of
cyber-attacks, since some cases substation
automation system are also connected to an external
network (e.g. office). So apart from protecting the
substation automation networkagainst unauthorized
access it is important that all deviceson this network
including POW controllers arealso robust with respect
to handling a possible attack. IEEE 1686 defines the
functions and features to be provided in intelligent
electronic devices (IEDs) to accommodate critical
260

April 8-10
2015

infrastructure protection programs. Some of the

measures to achieve robustness with respect to cyber
security include,
Port handling: Declaration of the ports which are used
by the IED, so that all other ports can be blocked by
the firewall.
User roles and accounts: Support for the creation of
user roles and accounts, so that only authorized users
can access the device for intended purposes.
generation of point-on-wave (POW) controllers,

designed on modern hardware and software
architecture, can offer additional benefits to the user,
as explained in this paper. These state-of-the-art
features will increase the value of the point-on-wave
controller to the user at no extra effort.
BIBLIOGRAPHY
1.
CIGRE Working Group 13.07: Controlled

Switching of HVAC Circuit Breakers: Guide for
Application Lines, Reactors, Capacitors,
Transformers. 1st part (LECTRA, No. 183,
April 1999, pp. 43-73).
2.
Stefan Meier: Enabling digital substations

(ABB Review, 4/2014, pp. 6-10).
3.
Hubert Kirrmann, Peter Rietmann, Steven

Kunsman: Standard Network Redundancy
Using IEC 62439 (PAC World, Autumn 2008,
pp. 38-44).
Certificate handling: Certificates for identification of

the devices while establishing the communication
so that only registered devices are in the network.
Encryption: For establishing the confidentiality of
the data transferred between devices,also for storing
user information.
8.
CONCLUSIONS
The basic technique of controlled switching has been

wellestablished for more than two decades. The latest
261

April 8-10
2015

Improve Transmission Fault Location and Distance

Protection Using Accurate Line Parameters
Hugo E. Prado-Flix
Vctor H. Serna-Reyna
Comisin Federal de Electricidad
Mangapathirao V. Mynam Marcos Donolo

Armando Guzmn
Schweitzer Engineering Laboratories, Inc.
ABSTRACT
Accurate positive- and zero-sequence line
impedances are critical for the performance of
distance relaying and fault location. This paper
discusses a method for calculating positive- and
zero-sequence line impedances under singlepole open conditions. Time-synchronized
measurements from both terminals of the line are
captured when one pole of either terminal is open.
This operational condition provides an unbalance
in the three-phase network, which is used to
measure the line impedances. The performance
of the method has been verified using
Electromagnetic Transients Program (EMTP)
simulations and field data provided by Comisin
Federal de Electricidad (CFE). The paper also
discusses existing methods to calculate line
parameters and provides a sensitivity analysis of
these methods. The paper presents an approach
where the calculated impedances can be
compared with the relay impedance settings and
an alarm sent if the difference is greater than a
predetermined threshold.
1.
and [2] use time-synchronized voltage and current

measurements from both terminals of the
transmission line to measure the impedance during
normal operating conditions.
Section II of this paper discusses the errors in line
parameters and their effect on distance protection
and fault location. Section III discusses classical
methods for determining positive- and zero-sequence
line impedances. Section IV discusses a method to
estimate the line parameters during normal and
single-pole open (SPO) operating conditions using
time-synchronized measurements. Section V
discusses the validation of the SPO-based method
using Electromagnetic Transients Program (EMTP)
simulations, and Section VI uses field data from a
Comisin Federal de Electricidad (CFE) 400 kV
network to estimate line parameters. Section VII
discusses a proposal to monitor the line impedances
during SPO conditions and send an alarm when the
estimated impedance and the relay settings differ
significantly.
2.
INTRODUCTION
Accurate transmission line parameters are critical for

many impedance-based applications, including
distance protection and fault location. Line
parameters are typically calculated based on tower
configuration and conductor physical properties.
Several factors affect the accuracy of the
calculations, including temperature and ground
resistivity. Some utilities conduct field tests to
measure the parameters of transmission lines. The
test process includes an outage of the line. The line
is shorted to ground at the remote end and energized
with a known source at the local end. References [1]
ERRORS IN LINE PARAMETERS AND

THEIR EFFECT ON DISTANCE
PROTECTION AND FAULT LOCATION
To show how errors in transmission line parameters

affect distance element and fault locating accuracy,
we first review the equations used by these
algorithms. Then we show the worst-case estimates
of the distance to the fault that include impedance
setting errors for a given fault condition.
A. Distance Protection
Distance-based line protection uses positive- and
zerosequence impedances along with voltage and
current measurements to determine if faults are inside
the protection zone. One method to estimate the
262

April 8-10
2015

distance to the fault for A-phase-to-ground faults

by mho elements is provided in (1). phase-to-ground
faults by mho elements is provided in (1).
Real (VaVpol*)
mAG =
....(1)
Real (Z1L(Ia k0IG)Vpol*)
where:
Va is the faulted phase voltage.
Vpol is the polarizing quantity.
Ia is the faulted phase current.
This method is typically used in line differential

relays, which have access to remote currents for
differential protection purposes. The accuracy of
this method is not influenced by system
nonhomogeneity and fault resistance; however, errors
in zero- and positive-sequence line impedances
influence the accuracy of the fault location. If
mutually coupled lines are present, zero-sequence
mutual impedance also affects the accuracy of the
estimation.
D. Multi-Ended Fault Location Using

Remote Terminal Currents and Voltages
IG is the residual current.

Z0L is the zero-sequence line impedance.
Z0LZ1L
k0 is the zero-sequence compensation factor
3Z1L
This fault location method uses voltages and

currents from the local and remote terminals [4] [5]. It
uses (4) to obtain the fault location for unbalanced
faults.
If the computed value mAG is less than the relay

reach setting, the relay declares an in-zone fault.
V2LV2R+12RZ1L)
FL_UNBME = Real
12TZ1L
B.
where:
Z1L is the positive-sequence line impedance.
Single-Ended Fault Location
This fault location method uses local voltages and

currents. Equation (2) shows the fault location
equation for an A-phase to-ground fault.
Imag (Va12*)
FL_AGSE =
Imag (Z1L(Ia+k0IG)12*)
....(2)
where:
I2 is the negative-sequence current.
The accuracy of the fault location depends on the
accuracy of the line parameter settings, fault
resistance, system nonhomogeneity, and zerosequence mutual impedance (if mutually coupled
lines are present).
C. Multi-Ended Fault Location Using

Remote Terminal Currents
This fault location method uses voltages and currents
from the local terminal of the line along with currents
from the remote terminal(s). Equation (3) shows the
fault location equation for A-phase-to-ground faults
[3].
Imag (Va12T*)
FL_AGMEI = ....(3)
Imag (Z1L(Ia+k0IG)12T*)
where:
I2T is the sum of the local and remote negativesequence currents.
....(4)
V2L is the negative-sequence voltage at the local

terminal.
V2R and I2R are the negative-sequence voltage and
current at the remote terminal, respectively.
This method can be used in line differential relays,
which have access to remote currents and voltages.
The accuracy of this method is not influenced by
the system homogeneity and fault resistance. A key
benefit of this method is that it is independent of
Z0L. Therefore, the actual fault location reported by
the line crew provides feedback on the accuracy of
positive-sequence line impedance.
E.
Distance to Fault Estimation Errors
We use (1) to study the effect of line parameter errors

on the distance to the fault. We represent the relay
setting Z1L as follows:
Z1L = Z1 + Z1
....(5)
where:
Z1 is the actual positive-sequence impedance value.
Z1 is the error.
To obtain the expression for Z0L, we replace 1 with
0 in (5). Fig. 1 shows the actual impedance Z1 and
the relay setting Z1L with an error with fixed
magnitude and variable phase angle.
263

April 8-10
2015

Fig. 1: Relay setting represented as the sum of the actual

impedance and setting error
Using (1) with Z1L and Z0L settings, Z1 equal to 3

percent, Z0 equal to 10 percent, and simulation results
for an A-phase-to-ground fault with zero fault
resistance at 80 percent of the line (see the appendix
for the system details), we find the minimum value
of m to be 0.73 of the line length. Proceeding in the
same way, we obtain a maximum value of m of 0.86
of the line length. Fig. 2 shows the minimum values
of m for all the combinations of errors in Z0L and
Z1L between 0 and 25 percent. Fig. 3 shows the
maximum values of m for all the combinations of
errors in Z0L and Z1L between 0 and 25 percent.
Fig. 3: Maximum values of m as a function of the error in Z1L and

Z0L for a fault at 80 percent of the line
The multi-ended fault location method that uses

voltages and currents from the remote terminal does
not depend on Z0L. Typically, Z0L has greater errors
than Z1L; therefore, this method provides more
accurate distance-to-fault estimates than the other
methods. Fig. 4 shows the effect of errors in Z1L on
the fault location estimation of this method.
These values are true for the mho element m

calculation, the single-ended fault location method,
and the multi-ended fault location method that uses
remote currents only.
Fig. 4: Maximum and minimum m values due to errors in Z1L for

the multiended fault location method
3.
METHODS TO DETERMINE LINE

PARAMETERS
Existing methods to determine the line parameters

can be categorized as follows:
Fig. 2: Minimum values of m as a function of the error in Z1L and
Z0L for a fault at 80 percent of the line
264
Parameter calculations using line constants

programs.

April 8-10
2015

Parameter measurements using signal injection

equipment.
Parameter estimation using time-synchronized

measurements (this method is discussed in
Section IV).
A. Line Constants Programs
Fig. 6: Test setup to measure the line parameters
Typically, line parameters are computed using line

constants programs, which are widely available.
These programs use tower configuration and
conductor properties to calculate the parameters.
Ground resistivity is one of the variables that is
difficult to determine. Ground resistivity, which
affects the resistance of the return path for the fault
current back to the substation ground, depends on
the terrain and weather. References [6] and [7] discuss
the impact of ground resistivity on the estimation of
the zero-sequence line impedance. Fig. 5 shows the
effect of ground resistivity on the zero-sequence
impedance of a typical 400 kV line.
At the other end of the transmission line, signals are

injected and voltage and current measurements are
taken to determine the line impedances. This method
requires three phase-to-phase impedance
measurements (Zab, Zbc, and Zca), along with three
phase-to-ground impedance measurements (Zag,
Zbg, and Zcg) and a zero-sequence impedance
measurement (Z0g) [8].
Positive-sequence impedance (Z1m) is computed
from the measurements using (6).
(Zab+Zbc+Zca)
Zlm =
....(6)
3
Zero-sequence impedance (Z0m) is computed using
(7) and (8).
(Zag+Zbg+Zcg)
Ze = Zlm
....(7)
3
(Zlm+3Ze)+Z0g
Z0m =
2
....(8)
It is important to note that these measurements do

not include errors in voltage transformers (VTs) and
current transformers (CTs) that still affect the
performance of the distance protection and fault
location schemes.
4.
Fig. 5: Effect of ground resistivity on zero-sequence impedance
B.
Signal Injection Methods for Measuring

Line Parameters
Signal injection is an option that utilities have to

measure line parameters; it requires a line outage and
an adequate power source.
The following procedure was used at CFE to
determine the parameters of lines where distance
protection schemes showed overreach problems.
As Fig. 6 shows, all three phases and ground
conductors (if present) are shorted and connected
to ground at one end of the transmission line.
ESTIMATING LINE PARAMETERS

USING TIME-SYNCHRONIZED
MEASUREMENTS
Time-synchronized measurements are available as

synchrophasors and as time-synchronized samples
of instantaneous signals. Time-synchronized
sampling is the mechanism where the data acquisition
in protective relays or digital fault recorders (DFRs)
acquires voltage or current samples with respect to
an absolute time reference, typically Global
Positioning System (GPS). Synchronized
measurements allow us to perform mathematical
operations on quantities measured at different
locations in the power system, in this case,
measurements from both terminals of the line.
265

April 8-10
2015

A. PI Model Approach
B.
A common approach to measure positive-sequence

line impedance uses a PI equivalent model of the
transmission line [9]. Fig. 7 shows the positivesequence PI equivalent of the line, and (9) estimates
the positive-sequence impedance.
The method discussed in the previous subsection

provides an estimate of the positive-sequence
impedance only. Unbalanced system conditions are
required to calculate the zero-sequence impedance.
Reference [7] discusses calculating the positive- and
zero-sequence impedances during ground faults; this
approach requires the actual fault location and the
fault voltages and currents at both terminals to
calculate the positive- and zero-sequence
impedances. Traveling wave fault location devices
are used for providing accurate fault location within
a tower span. Therefore, fault location results from
the traveling wave approach along with the voltages
and currents from both terminals of the transmission
line can be used to compute line parameters.
Z = R + jX =
Vs2 Vr2
Is VrIr Vs
....(9)
Fig. 7: PI equivalent model of the transmission line
The load angle between Vs and Vr and the CT and

VT errors have an important effect on the quality of
the impedance estimation in (9). Fig. 8 shows that for
small phase angle differences between Vs and Vr
(load angle less than 5 degrees), small CT and VT
errors are greatly amplified. Fig. 8 also shows that
CT and VT accuracies have a great impact on the
impedance estimation.
Single-Pole Open Method
Here, we propose calculating the positive- and

zerosequence impedances of the line while one
pole is open at only one end of the line. Fig. 9
shows a two-terminal system with A-phase open at
Terminal X.
For angles greater than 5 degrees, the calculation

error decreases rapidly. We show an example of this
problem in Section V, Subsection B.
Fig. 9: Two-terminal system with SPO condition
Fig. 8: Estimation error as a function of the angle between Vs

and Vr for three different CT and VT error ratings for the
PI model approach
This method can be used on transmission systems

with single-pole breaker control mechanisms and
line-side VTs [10] [11]. Fig. 10 shows the sequence
network for the line shown in Fig. 9. As we know,
the positive- and negativesequence line impedances
of transmission lines are the same. We selected the
negative-sequence network to calculate the negativesequence impedance because the effect of charging
current on the calculation method is lower in the
negativesequence network than in the positivesequence network when the phase is open at only
one end.
266

April 8-10
2015

As with the PI model approach, the load angle and

the CT and VT errors have an important effect on
the quality of the impedance estimation in (10) and (11).
Fig. 11 shows the positive-sequence impedance
calculation error, computed with (10) for the SPO
method as a function of the CT and VT errors and
the load angle.
This method is not recommended for lightly loaded
conditions. The zero-sequence mutual impedance in
mutually coupled line configurations affects the
estimation of zerosequence line impedance using the
SPO method.
Fig. 10: Symmetrical component network representation of an

SPO condition at Terminal X.
Equations (10) and (11) show the positive-, negative, and zero-sequence impedance calculations.
Z1L = (V2yV2x) / I2y = Z2L
....(10)
Z0L = (V0yV0x) / I0y
....(11)
Fig. 11: Estimation error in the positive-sequence line impedance
as a function of the load angle for two different CT and VT
error ratings for the SPO method.
where:
V2x is the negative-sequence voltage calculated at
Terminal X.
V2y is the negative-sequence voltage calculated at
Terminal Y.
I2y is the negative-sequence current calculated at
Terminal Y.
V0x is the zero-sequence voltage calculated at
Terminal X.
5.
EMTP SIMULATION RESULTS
We modeled a two-source one-line system to verify

the performance of the SPO method. The modeled
252-kilometer transposed line is part of the CFE 400
kV transmission network between the MZD and TPD
substations. Fig. 12 shows the one-line diagram with
relays connected to both terminals.
V0y is the zero-sequence voltage calculated at

Terminal Y.
I0y is the zero-sequence current calculated at
Terminal Y.
The SPO method depends on the negative-sequence
voltage difference between the two terminals for the
negativesequence impedance estimation and on the
zero-sequence voltage difference for the zerosequence impedance estimation.
267
Fig. 12: Test system with line relays connected to GPS

April 8-10
2015

The power system was modeled in a Real Time Digital

Simulator, which runs EMTP, along with CTs and
coupling capacitor voltage transformers (CCVTs). The
outputs of the CTs and CCVTs are connected to the
relays protecting the 252-kilometer transmission line.
The relays at both terminals are configured to trigger
event reports when line faults occur. Additionally,
both relays are configured to stream synchrophasor
data to a phasor data concentrator with archiving
capability.
A. SPO Method Using Time-Synchronized

Instantaneous Samples
An A-phase-to-ground fault was simulated at 242
kilometers from the MZD terminal. Fig. 13 shows
the time-aligned event reports from the MZD and
TPD terminals. The relay at TPD detected the fault
and tripped the breaker half a cycle earlier than the
relay at MZD.
Fig. 14: Signal processing to compute Z1L and Z0L
per cycle. A cosine filter is then used to extract the

fundamental quantity of the signal and filter out dc
and harmonics. Processing in the highlighted section
of Fig. 14 is not required when using filtered event
reports. The cosine-filtered signal is then used to
construct the phasors.
We use the logic shown in Fig. 15 to identify local
SPO conditions while the remote terminal is closed.
Equations (10) and (11) are used to calculate the
line impedances based on the data corresponding to
this condition. Fig. 16 shows the comparison of the
estimated positivesequence impedance during the
SPO condition with the line impedance calculated
by the line constants calculation (LCC) program. Fig.
17 shows the same comparison for zerosequence
impedance.
Fig. 18 shows the real-time zero-sequence resistance
and reactance calculations (implemented using relay
programmable logic) performed by the relay at the
MZD terminal using time-synchronized local and
remote voltages and currents [12].
Fig. 13: Faulted phase voltages and currents at both terminals for
a fault at 242 kilometers from MZD
For this fault, the TPD terminal recloses prior to the

MZD terminal after the 1.2-second open-interval
timer expires. The MZD terminal has an SPO
condition with the TPD terminal closed. Voltages and
currents from both terminals in this period of an SPO
condition are used for computation of the positiveand zero-sequence line impedances. Data are
postprocessed to compute the positive- and zerosequence line impedances. Fig. 14 shows the signal
processing flow.
Table 1 shows the average line impedance results

estimated during SPO conditions using timesynchronized samples.
Event reports are retrieved from both terminals and

timealigned to adjust for different trigger times. If
the event reports are based on a fixed sampling rate
(samples per second), the aligned data are processed
using a low-pass filter and resampled to 32 samples
268
Fig. 15: Logic to detect SPO condition with the remote terminal
closed

April 8-10
2015

Table 1: Line Impedances Estimated With The Lcc

Program And During Spo Conditions Using Timesynchronized Samples
Method
Positive
Sequence
(primary ohms)
Zero
Sequence
(primary ohms)
LCC program
7.2 + j97.32
86.52 + j323.4
SPO
8.16 + j97.27
80.59 + j323.3
B. Synchrophasor-Based SPO Method

Synchrophasor measurements were archived during
the simulation. Fig. 19 shows the faulted phase
current and voltage magnitudes, during pre-fault,
fault, and post-fault.
Fig. 16: Positive-sequence line impedance comparison between

the LCC program and SPO method
Fig. 19: Synchrophasor measurements from both terminals for a

trip and reclose sequence (MZD is shown by the red trace; TPD
by the blue)
Fig. 17: Zero-sequence line impedance comparison between the
LCC program and SPO method
Table 2 shows the line impedance calculated during

the normal load conditions using the PI model
approach and during the SPO condition using
synchrophasors in primary ohms.
Table 2: Line Impedances Estimated Using
Synchrophasor Measurements
Fig. 18: Real-time zero-sequence impedance calculations
269
Method
Positive
Sequence
(primary ohms)
Zero
Sequence
(primary ohms)
PI model
6.96 + j 95.59
NA
SPO
5.96 + j96.54
86.11 + j324.7

April 8-10
2015
6.

FIELD RESULTS
A. Signal Injection-Based Testing at CFE

We compared line parameters obtained using the
signal injection method and the LCC program on four
transmission lines.
(1) GDU-CLN 6.5-Kilometer 230 kV Line
The first line we studied was the 6.5-kilometer 230
kV line from GDU to CLN, which is an underground
cable. Fig. 20 shows the measured (Z0m and Z1m)
and computed (Z0lcc and Z1lcc) values.
Fig. 21: Measured and computed positive- and zero-sequence

impedances for a 14.8-kilometer 69 kV line
(3) APT-AGT 36-Kilometer 230 kV Line

The third line we studied was the 36-kilometer 230
kV line from APT to AGT, which is an overhead line.
Fig. 22 shows the measured and computed positiveand zero-sequence impedance values.
Fig. 20: Measured (Z0m and Z1m) and computed (Z0lcc and
Z1lcc) values for a 6.5-kilometer 230 kV line. Note the disparity
between Z0m and Z0lcc
If we use the measured impedances to compute

currents and voltages for an A-phase-to-ground fault
at 80 percent of the line length, a distance function
using the impedance values from the LCC program
will see the fault at 95 percent of the line.
(2) GDU-SCU 14.8-Kilometer 69 kV Line
The second line we studied was the 14.8-kilometer
69 kV line from GDU to SCU, which is part
underground cable and part overhead line. Fig. 21
shows the measured and computed positive- and
zero-sequence impedance values.
In this case, the difference between the measured
and the computed values of the line impedances
causes a fault at 80 percent of the line to look like a
fault at 67 percent, which causes the distance
element to overreach.

impedance for a 36-kilometer 230 kV line. Note that while the X/R
ratios are very accurate, there is a large difference in magnitude
In this case, the differences between the measured

and computed positive- and zero-sequence
impedance values are significant. These differences
cause a fault at 80 percent of the line to look like a
fault at 59 percent, which implies that the relay will
overreach by about 21 percent. In this case, CFE
modified the relay settings according to the
measurements.
270

April 8-10
2015

(4) AGS-AGN 5.4-Kilometer 115 kV Line

Finally, we studied the 5.4-kilometer 115 kV line
between AGS and AGN, which is an overhead line.
Fig. 23 shows the measured and computed positiveand zero-sequence impedance values.
Fig. 25 shows the line resistance and reactance

calculated using (9) during the 24-hour period.

impedances for the 5.4-kilometer 115 kV line
Fig. 25: Estimated line resistance and reactance during a

24-hour period
In this case, the difference between the measured

and the computed values of the positive- and zerosequence impedances causes a fault at 80 percent of
the line to look like a fault at 88 percent.
We can summarize the previously described results
as follows:
z
Signal injection is a viable option for utility

personnel to measure the line parameters.
Differences between Z1m and Z1lcc are smaller

than the differences between Z0lcc and Z0m.
B.
PI Model Approach to Estimate Z1L
Synchrophasor measurements were archived from

both ends of a 383-kilometer line during a period of
24 hours. Fig. 24 shows the voltage angle difference
between the line ends.
The LCC-calculated positive-sequence impedance for

this line is 11.6 + j133.4 Averaging the Z1L values
obtained during the 24-hour period, the PI model
Z1L is 10.99 + j143.8 which is within 10 percent
of the LCC calculated value and provides an
independent verification of the positive-sequence
line impedance.
We classified the estimations in clusters based on
the angle difference between the terminals of the
line. Fig. 26 shows the impedance variation with
respect to the line loading angle. In this case, the
top right cluster corresponds to a 35-degree
difference. Clusters to the left and down correspond
Fig. 24: Positive-sequence voltage angle difference between the

ends of the transmission line during a 24-hour period.
Fig. 26: Line impedance estimation variation with line loading
271

April 8-10
2015

to decreasing angle values, with the cluster at the

bottom left corresponding to a 15-degree difference.
C. Line Parameter Calculation Using the

SPO Method
Events recorded during a C-phase-to-ground fault (at
242 kilometers from MZD) on a line between MZD
and TPD were used to verify the line parameter
estimation using the SPO method. Fig. 27 and Fig.
28 show the events captured at the MZD and TPD
terminals.
before the MZD terminal. CFE uses a sequential

reclosing system that monitors the generation at the
two line terminals and recloses the terminal with the
lowest generation first, which explains the longer
reclosing delay at MZD. The data in the SPO window
are used to compute the positive- and zero-sequence
impedances of the line. Line reactors are in operation
at both terminals during this event; therefore, shunt
reactor compensation is included in the SPO
calculation. Equation (12) shows the provided
compensation.
Vp
Ipcompensated = Ip
....(12)
Xr
where:
Ip is the phase current.
Vp is the phase voltage.
Xr is the reactor impedance.
The compensated currents are used in (10) and (11)
to compute the positive- and zero-sequence
impedances of the line.
The line impedance is computed for each sample in
the SPO data window. Fig. 29 shows the positivesequence impedance of the line in the impedance
plane as calculated by the LCC program and the line
impedance estimated using the SPO method.
Fig. 27: Event recorded at MZD terminal
Fig. 28: Event recorded at TPD terminal
The captured events include the pre-fault load

condition, C-phase-to-ground fault, C-phase pole
open period, and successful reclose at both terminals.
In this event, the TPD terminal recloses 22 cycles
272
Fig. 29: Positive-sequence line impedance comparison between

the LCC and the SPO methods

April 8-10
2015

Fig. 30 shows the comparison of the zero-sequence

line impedance using the SPO method versus the LCC
program method.
7.
FEEDBACK TO DISTANCE AND FAULT

LOCATION FUNCTIONS
The event reports from the MZD-TPD line (shown in

Fig. 27 and Fig. 28) provide voltages and currents
during the fault, during the phase open condition
(open poles at both terminals), and when one pole is
open at only one terminal (the condition that the
proposed SPO method requires).
After the first pole recloses and there is no fault, the
relay identifies the desired SPO condition, verifies
line loading, calculates the line sequence impedances,
and compares them with the relay impedance
settings. If the difference is greater than a
preestablished threshold, the relay asserts an alarm.
Referring to the MZD-TPD line, we have analyzed
only one event that shows differences of 6 percent
in Z1L and 16 percent in Z0L, which translates to a
difference of less than 2 percent in the m calculation
(Fig. 31). We will analyze more events for faults at
other locations and continue monitoring the
estimations of Z1L and Z0L before changing the
impedance settings.
Fig. 30: Zero-sequence line impedance comparison between the

LCC and the SPO methods.
Table 3 shows the values of the estimated

line impedances during the SPO condition and
the calculated impedances using the LCC
program.
Table 3: Estimated Line Impedances with the LCC
Program and During SPO Conditions
Method
Positive
Sequence
(primary ohms)
Zero
Sequence
(primary ohms)
LCC program
7.2 +j97.32
86.52 + j323.4
SPO
8.45 +j91.66
73.42+j269.45
Fig. 31 shows the correlation of the faulted phase

currents at TPD and MZD with the distance element
m calculation of the MZD relay during the fault using
the LCC and SPO impedance estimates from Table
3. Fig. 31 shows that, in this case, the impedance
reach setting margins for Zone 1 and Zone 2
accommodate the differences in the m estimates
from the LCC and SPO methods.
273
Fig. 31: Calculation of m during the phase-to-ground fault

April 8-10
2015

For the event that we analyzed, the margins of the

present relay settings accommodate the existing line
parameter, instrument transformer, and relay errors.
8.
CONCLUSION
Relay engineers calculate transmission line positiveand zero-sequence impedances that are needed to
set the relay distance and fault location functions
using conductor properties, line geometry, and
ground resistivity as inputs to line parameter
calculation programs.
In some cases, they apply reduced primary voltages
at one end of the line while the remote end is
connected to ground to measure these impedances
and verify the results from the line parameter
calculation program. While this approach provides
accurate impedance values, it does not consider VT
and CT errors and that impedance value
measurements are difficult to obtain.
Fig. 32: Two-source power system model
Voltages and currents at both terminals of the line

were computed for an A-phase-to-ground fault with
zero fault resistance at 80 percent of the line. Table
IV shows the secondary phase voltages and currents
during this fault condition.
Table 4: Fault Voltages and Currents for a Fault at
80 Percent from the Sending End
Voltages and
Currents
Synchrophasor measurements during normal

operating conditions provide information about the
positive-sequence impedance but not the zerosequence impedance.
Sending
End
Receiving
End
A-Phase Voltage
44.914.81 V
8.024.59 V
B-Phase Voltage
65.68110.64 V
74.11126.34 V
C-Phase Voltage
65.29126.17 V
76.33134.15 V
A-Phase Current
7.469.93 A
5.8182.95 A
B-Phase Current
1.1388.76 A
1.1391.23 A
C-Phase Current
0.31141.44 A
0.3138.56 A
This paper presents a method to estimate positiveand zero-sequence impedances, including VT and CT
errors, without the need for signal injection. This
method requires line-side VTs, breakers with singlepole trip/close mechanisms, voltage and current
synchronized measurements, a load angle greater
than 5 degrees, and an operating condition where
one pole is open at one of the terminals while all of
the poles at the remote terminal remain closed. We
can use this method to verify the line parameters
obtained with line parameter calculation programs
and signal injection methods.
Acknowledgement
1.
M. Grobler and R. Naidoo, Determining

Transmission Line Parameters From GPS TimeStamped Data, proceedings of the 32nd Annual
Conference on IEEE Industrial Electronics, Paris,
France, November 2006.
Relays can detect single-pole open conditions, perform

impedance estimations, and provide alarms when the
difference between the estimated values and the relay
settings is greater than a predetermined threshold.
2.
D. Shi, D. J. Tylavsky, N. Logic, and K. M.

Koellner, Identification of Short TransmissionLine Parameters From Synchrophasor
Measurements, proceedings of the 40th North
American Power Symposium, Calgary, Canada,
September 2008.
3.
B. Kasztenny, B. Le, and N. Fischer, A New

Multiterminal Fault Location Algorithm
Embedded in Line Current Differential Relays,
proceedings of the 11th International
Conference on Developments in Power System
Protection, Birmingham, UK, April 2012.
9.
The authors would like to thank Mr. Raj Suyambu

for verifying the proposed method to estimate line
parameters.
REFERENCES
APPENDIX
The power system model in Fig. 32 (including the

shown positive- and zero-sequence secondary
impedance values) was used to analyze the effect of
errors in line impedances on the performance of mho
distance element and fault location functions in line
relays.
274

April 8-10
2015

4.
D. A. Tziouvaras, J. Roberts, and G. Benmouyal,

New Multi-Ended Fault Location Design for
Two- or Three-Terminal Lines, proceedings of
the 7th International Conference on
Developments in Power System Protection,
Amsterdam, Netherlands, April 2001.
5.
Y. Gong, M. Mynam, A. Guzmn, G. Benmouyal,

and B. Shulim, Automated Fault Location
System for Nonhomogeneous Transmission
Networks, proceedings of the 65th Annual
Conference for Protective Relay Engineers,
College Station, TX, April 2012.
6.
P. M. Anderson, Power System Protection. IEEE

Press, New York, NY, 1999.
7.
A. Amberg, A. Rangel, and G. Smelich,

Validating Transmission Line Impedances
Using Known Event Data, proceedings of the
65th Annual Conference for Protective Relay
Engineers, College Station, TX, April 2012.
8.
R. Luxenburger and W. de Villiers, Calculation

and Verification of Distance Protection Settings
Based on Line Impedance Measurements,
proceedings of the 4th Annual Protection,
Automation and Control World Conference,
Dublin, Ireland, June 2013.
9.
R. Abboud, W. F. Soares, and F. Goldman,

Challenges and Solutions in the Protection of
a Long Line in the Furnas System, proceedings
of the 32nd Annual Western Protective Relay
10. F. Calero and D. Hou, Practical Considerations

for Single-Pole-Trip Line-Protection Schemes,
proceedings of the 31st Annual Western
Protective Relay Conference, Spokane, WA,
October 2004.
11. V. H. Serna Reyna, J. C. Rivera Velzquez, H. E.
Prado Flix, H. J. Altuve Ferrer, D. Snchez
Escobedo, and J. Gallegos Guerrero,
Transmission Line Single-Pole Tripping: Field
Experience in the Western Transmission Area
of Mexico, proceedings of the 37th Annual
Western Protective Relay Conference, Spokane,
WA, October 2010.
12. A. Guzmn, V. Mynam, and G. Zweigle, Backup
Transmission Line Protection for Ground Faults
and Power Swing Detection Using
Synchrophasors, proceedings of the 34th

Annual Western Protective Relay Conference,
Spokane, WA, October 2007.

AUTHORS
Hugo E. Prado-Flix received his BSEE in 1982 from
the National Polytechnic Institute, Mexico. In 1983,
he joined Comisin Federal de Electricidad (CFE),
where he was a transmission inspector of the Colima
Transmission Subarea from 1983 to 1985 and the
Michoacn Transmission Subarea from 1985 to
1987. He was also head of the Protection Office of
the Nayarit Transmission Subarea from 1987 to 2002
and head of the Jalisco Transmission Subarea from
2002 to 2005. From 2005 until 2007, he was head
of the Protection Department of the Jalisco
Transmission Subarea. Since 2007, he has worked
as head of the Protection and Metering Office of the
CFE Western Transmission Area, where he supervises
the installation and commissioning of protection,
control, and metering panels.
Vctor H. Serna-Reyna received his BS in mechanical
and electrical engineering in 1992 and an MSc in
electrical engineering in 1997 from the University
of Guadalajara, Mexico. In 1997, he joined Comisin
Federal de Electricidad (CFE). Since 2007, he has
worked as head of the Studies Department of the
CFE Western Transmission Area. His areas of interest
are power system analysis and protection.
Mangapathirao V. Mynam received his MSEE from
the University of Idaho in 2003 and his BE in
electrical and electronics engineering from Andhra
University College of Engineering, India, in 2000.
He joined Schweitzer Engineering Laboratories, Inc.
(SEL) in 2003 as an associate protection engineer
in the engineering services division. He is
presently working as a senior research engineer in
SEL research and development. He was selected to
participate in the U.S. National Academy of
Engineering (NAE) 15th Annual U.S. Frontiers of
Engineering Symposium. He is a senior member of
IEEE.
Marcos Donolo received his BSEE from Universidad
Nacional de Ro Cuarto, Argentina, in 2000, and his
masters degree in electrical engineering (2002), his
masters degree in mathematics (2005), and his Ph.D.
275

April 8-10
2015

in electrical engineering (2006) from the Virginia

Polytechnic Institute and State University. Since
2006, he has been with Schweitzer Engineering
Laboratories, Inc., where he is presently a lead
research engineer. He is a senior member of IEEE.
Armando Guzmn received his BSEE with honors
from Guadalajara Autonomous University (UAG),
Mexico. He received a diploma in fiberoptics
engineering from Monterrey Institute of Technology
and Advanced Studies (ITESM), Mexico, and his
MSEE from the University of Idaho, USA. He served
as regional supervisor of the Protection Department

in the Western Transmission Region of the Federal
Electricity Commission (the Mexican electrical utility
company) in Guadalajara, Mexico, for 13 years. He
lectured at UAG and the University of Idaho in power
system protection and power system stability. Since
1993, he has been with Schweitzer Engineering
Laboratories, Inc., in Pullman, Washington, where
he is a fellow research engineer. He holds numerous
patents in power system protection and metering.
He is a senior member of IEEE.
276

April 8-10
2015

Fault Current Limiter Selection Considerations

for Utility Engineers
Suraj Rengarajan
Vikas Dabeer
K. Tekletsadik
P. Lubicki
J. Ludlum
P. Murphy
S. Kodle
A.Sinha
Applied Materials India Pvt.Ltd., India
SUMMARY
The need for Fault Current Limiters are driven by
rise in the fault current levels as the energy
demand increases and more generation,
distribution and distributed energy generation
such as wind, solar, gas powered generation and
diesel generators are added to an already
overburdened system.
Several Fault Current Limiter (FCL) technologies
have matured from R&D and demonstration
projects into commercially available systems
worldwide. So far, the technical knowledge about
FCL systems, design parameters and impact
analysis on utility equipment is primarily
contained within the FCL manufacturers technical
expertise. The time has come to further
disseminate the technical knowledge to utility
engineers so that they can design fault current
management systems with the technical expertise
that enables them to specify FCLs for their
applications. Applied Materials is developing
Superconducting Fault Current Limiters (SCFCL)
and Solid State Fault Current Limiters (SSFCL)
for transmission and distribution voltage levels.
Both FCL technologies are being subjected to
testing and in-grid demonstration.
This paper is aimed at describing how utility
engineers may consider selectinga FCL for a
specific location and application, based on the
authors experience. It is not intended as a manual
or guide to be used or relied upon without
independent testing or verification. Italso
discussesthree FCL demonstration projects
of Applied Materialsincluding SCFCLs
in California and New York, and a SSFCLin
Australia.
S. Nickerson
Keywords: Fault Current Limiter, Superconducting,

Solid State, FCL, SCFCL, SSFCL.
1.
INTRODUCTION
Despite having more than 225 GW of installed

generation capacity, India continues to be power
deficit. One of the major reasons for this situation is
the inadequate transmission capacity, not matching
the generation capacities and load requirements. With
the sole exception of captive power, cross country
transmission lines ferry every unit of the power
Generated in the country.Power evacuation is turning
out to be a bigger problem than power generation
for the country. Plants supplying electricity to state
electricity boards (SEBs) under long term power
purchase agreements (PPA), lost 1.93 billion units
of generation due to transmission capacity
bottlenecks.[1]
Fault current levels are increasingas an unwelcome
consequence of increased generation, transmission
systems development and more interconnected grid
to have higher grid stability and reliability. Many
utilities are experiencing the problem of astonishing
fault Current levels. It is necessary to bring down
the fault current levels to acceptable levels of
equipment rating for a safer and reliable operation
of the grid.Increasing rate of fault level causes
undesired consequences such as unacceptable
thermal electro-dynamic stresses, equipment damage,
and safety problem to personnel etc. Even a small
rise in fault level compared to equipment rating can
cause failures and reduce the life of expensive
equipments at substation.Fault current Limiters can
help the utilities to mitigate the fault current
problems.
As FCL technology transforms into a market-
277

April 8-10
2015

readysystem, the design and specification of the

systemtransitions from the manufacturer to the
utility engineers. Applied Materials is taking a lead
in this transformation and isdeveloping an FCL
family of systems based on Superconducting
(SCFCL)andsolid state (SSFCL) technologies. The
main objective of this publication is to provide
considerations for utility engineers in selecting the
right FCL solution for a specific location and
application. The FCL component has been
developed for PSCAD, DigSilent, PSS/E, EMTP, and
PSpice circuit simulation software packages. In
general, any of the available software packages can
be used with no or minimal loss of accuracy.
rating results with benefits of improved transient

stability and voltage stability that improves the
system resilience to power outages. Figure 1 shows
how the FCL impacts the future generation,
transmission and distribution systems by decoupling
the fault current constraint from the system design.As
the cost of FCLs continues to decrease and pilot
projects move into deployment, FCLs have the
potential to alter the way we design and plan the
electrical grid.
This paper will also discuss three FCL demonstration

projects of Applied Materialsincluding SCFCLs
inCalifornia and New York and a SSFCLin Australia.
Sharing our experiences in the design, manufacturing
and installation of FCL demonstration systems is one
of the main goals of this paper. These demonstration
units are installed and being evaluated over a period
of one-to-two years in order to demonstrate the
systemsreliability over longer-term operating
conditions.
2.
BENEFITS OF FAULT CURRENT

LIMITERS
The benefits of FCLs can be categorized based on

two system types. The first one is related to existing
systems where the main expected benefits of FCLs
are increased asset utilization, equipment upgrade
deferral, life extension, improved safety, improved
reliability and operational flexibility, enabling
interconnection of grids and avoiding major projects
like splitting buses and building new substations.
The second type of benefitis for new systems design
where a fault current constraint is removed from the
critical system design equation and all equipment
isdesigned with low fault currents. As an example, a
new system that uses FCLs could be designed for a
maximum fault current of 50% or less than its
equivalent old system. It can also mean power
system equipment manufacturers build cost-effective
and low-fault current rating equipment such as
breakers, low impedance transformers, compact
reactors, etc. With the FCLs new approach, a higher
quality of power with stable systems can be designed.
New systems can be designed for low fault current
Fig. 1: Impact of FCL on future Power System design
3.
FCL TECHNOLOGIES AND

APPLICATIONS
Applied Materials is developing two types of FCL

technologies, Superconducting (SCFCL) and Solid
State (SSFCL). The FCL systems are based on a
modular design platform that can be easily
configured to meet various customer needs based
on location, space availability and applications. The
SCFCL is primarily for transmission voltage levels
from 66 kV to 400 kV; whereas the SSFCL is targeting
distribution voltage levels of up to 66 kV.
The application of FCLs may coverall sizes of utilities
in their generation, transmission and distribution
systems. Independent power producers and industrial
customers also could benefit fromFCLs. Based on
the installation and configuration types, FCLs could
be installed in-line with a feeder, in a bus-tie
application or transformer neutral-to-ground
connections.
Figure 2 shows a typical bus-tie FCL application,
where it is primarily used to interconnect buses. This
278

April 8-10
2015

application can also be extended to system

interconnections of larger grids or substations to
increase power availability and system reliability. The
bus-tie application is also very useful for voltage sag
improvement, where the voltage drop at the unfaulted (healthy) section can be maintained to a value
that does not affect the customers connected to that
section. This application improves grid
interconnection and enables the addition of
renewable energy generation with minimal impact
on the existing system.
4.
INSTALLATIONS
Figure 4 shows a distribution SCFCL installed in Santa

Clara, CA. It is a 15 kV class, 1000 a 3 phase SCFCL
and uses an active Liquid Nitrogen Supply. This
system has been installed and running in the system
for the last 18 months and has operated as designed
to date. This technology could also be supplied with
an optionalbulk Liquid Nitrogen system which
reduces the overall system complexity and cost.
Reactors
Cryostat - Super
conductor Unit
Cryogenics Cooling system
Fig. 4: SCFCL at Silicon Valley Power, Santa Clara, California
Fig. 2: Typical Bus-tie FCL Application
The other type of FCL connection is in-line (shown

in Figure 3) with a feeder or transformer to limit the
fault current contribution from that line and protect
it from fault currents from the rest of the system.
Figure 5 shows a recent installation at Central Hudson

Knapps Corner Substation in Poughkeepsie, New
York. It is a 15 kV class neutral-to-ground SCFCL.
This SCFCL has already experienced several faults
at grid and performed as expected.
Fig. 5: SCFCL at Central Hudson, Poughkeepsie, New York
Fig. 3: Typical in-Lines FCL Application
279

April 8-10
2015

A neutralto-ground FCL is used where more than

80% of faults are phase-to-ground faults. In such
applications, using a neutral-to-ground FCL seems
to be a reasonable solution. It is understandable that
20% of faults that are phase-to-phase or balanced
three phase faults are not limited with the neutralto-ground FCL. The other concern with the
application of neutral-to-ground FCL system,
especially at higher voltage applications, is the
overvoltage in the healthy phases.
of the cost drivers for the FCL system they are

planning to buy. The major cost drivers are the load
current (IL) and the fault current reduction required
which is directly proportional to the voltage drop
across the FCL during fault (VFCL). It is therefore
possible to get a cost factor (CF) related to the rating
of the FCL as;
For example: for neutral-to-ground FCL with a 50%

current reduction, the overvoltage on the healthy
phases could go up to 1.32VP (32% overvoltage). This
may be acceptable at distribution systems but needs
to be checked for insulation coordination purposes
at transmission voltage levels.
This equation shows that ahigher the load current

and current reduction results in higher cost of the
FCL. The size and weight of the FCL is also linearly
proportional to the cost factor. Understanding these
factors helps the utility engineers to optimize the
FCL location and application to get a cost-effective
solution.
Figure 6 shows a Solid State FCL installed at grid.

This system is a 23 kV SSFCL built to limit fault
currents from up to 26 kA symmetrical rms (65 kA
peak) to as low as less than 6 A peak. This SSFCL is
designed to work for an application where the
required limited fault current varies between 6 A
peak (99.99% current reduction) to 40 kA peak (23%
current reduction). SSFCL also has an added
advantage of instant recovery for applications where
instant recovery is critical, such as systems that are
sensitive to voltage drops during recovery time.
CF = k1.IL.CR = k2.IL.VFCL = k2.IL.ILim.Xsh

The cost factor constants k1 and k2 are FCL-type
dependence and are expected to decrease as the FCL
technology matures and the cost of the FCL decreases
through materials performance improvements and
as the cost of volume manufacturing decreases.
6.
The most important system parameters required for

selecting a FCL are listed in Table 1. This table is
intended as an example of how to calculate the FCL
design parameters and provides a means to write a
specification or rating of the selected FCL.
As shown in Table 1, a utility engineer can calculate
the key system parameters for short circuit analysis.
Based on these calculated parameters one could
estimate the specifications or ratings for an FCL for
that application. The example above describes this
process for a 220 kV system. Since most short -circuit
analysis isdone with power system software
packages, the next section describes the best
approach a utility engineer could take to design an
FCL of his or her choice.
7.
Fig. 6: SSFCL at AUSNET, Australia
5.
COST DRIVERS
When looking for a FCL, engineers should be aware
SYSTEM REQUIREMENTS
FCL SIMULATION MODELS
This section addresses a key skill set fora utility

engineer in how to incorporate FCLs into their shortcircuit analysis. Applied Materials has developed
several simulation models for both SCFCL and
SSFCLsystems. These models went through extensive
revisions and were experimentally verified through
extensive testing at component and system levels.
280

April 8-10
2015

Table 1: FCL System Requirement Table
state units. Further simplification shows that the

Superconductor unit can be represented by a fixed
resistor with a value around 3 to 5 times the parallel
shunt reactor. Even further simplification shows that
a current limiting reactor model is accurate enough
that the simulation has no loss of accuracy. Utilities
that use frequency-based software, like PSS/E may
be able to use thesimplest model.
System Parameters - Provided by Utility

System Voltage - Line-to-Line, Vs
220
KV rms
Maximum Load Current, IL
1500
A rms
Prospective Fault Current, Ip
25
kA rms
Limited Fault Current, ILim
12.5
kA rms
Calculated System Parameters

System Short Circuit Impedance, 5.08
vs
Zs =
Ip.3
IpILim
Current Reduction, CR =
50
Ip
5.08
Voltage Drop Across FCL,

VFCL = Zsh.ILim
63.5
kV rms
Recovery time after Fault

is cleared
2.0 - 3.5 sec
Shunt Reactor Impedance,

CR
Zsh = Zs
1CR
Fault Current Limiter Rating

220 kV System Voltage, with 63.5 kV Voltage drop,
1500A, 5.08 Shunt Reactor, Limits 25 kA to 12.5
kA (50% fault Current Reduction) and recovers to its
normal operating mode within 3.5 seconds
Fig. 7: FCL Simulation Models
In addition to the in-house test lab for component

and module level testing, KEMA power test labs has
been used extensively to characterize the
performance properties of the FCL systems. FCLs
have undergone over seven weeks of KEMA testing,
exposing the system to thousands of high power
faults, and have performed per design specifications
and simulation results.
Figure 8 shows some of the basic simulation models
developed for FCL systems. These models start from
physics-based models that include transient electrical
and thermal analysis and progress to simplified
models with minimal or no loss of accuracy. Our inhouse design software and PSCAD use the physicsbased model. For some commercial software
packages, like DIgSilent, ramped and step
approximation functions are used to model the
transition resistance of the Superconductor and Solid
Figures 8 to 13 show simulation results for different

types of FCL models. In general, since the fault
current is much greater than the trigger current for
the FCL, the FCL circuit in either the SCFCL or SSFCL
introducessufficient impedance within 1- 2
ms.Meanwhile, most of the current transfers to the
current limiting shunt reactor. The difference
between the physics-based simulation model and
those approximations is negligible.
It is important to note that the way both SCFCLs
and SSFCLs are designed, the transient overvoltage
across the parallel connection of the shunt reactor
and the Superconductor or Solid State unit is
controlled with voltage control circuits and the
simulation models should be used with no concern
about the transient overvoltage. The impact of the
FCL on Transient Recovery Voltage for breakers is
therefore negligible.
281

April 8-10
2015

Figure 8 shows how the Superconductor senses fault

current, quenches, introduces high resistance and
current transfers to the parallel shunt reactor and
hence fault current is limited by the reactor. This
simulation model is based on a 220 kV transmission
SCFCL with a prospective fault current of 25 kA
(symmetrical rms) and designed to limit the fault
current to 12.5 kA, a 50% current reduction including
the first peak.
current flowing through the SS unit after it triggers

the unit and almost all the fault current transfersto
the parallel shunt reactor.Even with further
simplification using a series current limiting reactor,
as shown in Figure 11, there is no loss of accuracy
on the limited fault current waveform or amplitude.
Fig. 10: SSFCL Current limiting simulation
Fig. 8: SCFCL current limiting performance simulation using

superconductor physics and thermal model
Fig. 11: FCL Simulation using a series current limiting reactor
Fig. 9: SCFCL model using a reactor in parallel with a fixed

resistor
As seen on Figure 9, using a simplified parallel reactor

and resistor model misses the transient effects of
the Superconductor heating at the start of the fault
for around 1 to 2 ms. Neglecting this effect did not
affect the overall limited fault current waveform
shape of amplitude.
Figure 10 shows a simulation of a SSFCL using a
pre-set trigger current. This simulation model is based
on a SSFCL with a prospective fault current of 25
kA (symmetrical rms) and is designed to limit the
fault current to 12.5 kA, a 50% current reduction
including the first peak. Compared with the SCFCL,
SSFCL simulation is even more suitable for further
simplification because of negligible orextremely low
Figure 12 shows a comparison of four different

models described above. For the purpose of shortcircuit analysis, utility engineers can use even the
simplest model with negligible or no loss of accuracy.
The key takeaway here is the FCL knowledge is now
available to utility engineers in the form of a
simplified simulation model. Utility engineers can
design their own FCL and ask the manufactures to
make their selected system. It is as simple as selecting
the value of a series current limiting reactor.
One additional FCL performance parameter is the
recovery time after the fault clears. Recovery time
is mostly associated with SCFCL for obvious
reasons,as the superconductorheated and
transitioned to non-superconducting state by the
fault requires time to cool and return to the
superconducting state. The Applied Materials SCFCL
282

April 8-10
2015

is designed to recover within 2 to 3.5 seconds after

the fault is cleared. Figure13 shows thetypical
recovery time istakes asuperconductor unit to
recover ataround 2.2 seconds.
stability and protection issues, SSFCL appears to give

a clear advantage over SCFCL.
8.
CONCLUSION
As the cost of FCLs continues to decrease and pilot

projects move into deployment, FCLs have the
potential to alter the way we design and plan the
electrical grid. This smart technology can improve
safety and reliability, defer capital and offer
operational flexibility.Utility engineers may wish to
consider adding FCL technologies and systems to
the utility toolbox and using them to design a costeffective fault current management system
forexisting or future systems.
Fig. 12: A comparison of four different models described above
BIBLIOGRAPHY
1.
FICCI Report on Power Transmission The

Real Bottleneck, and The Financial Express
http://www.ficci.com/spdocument/20311/
power-transmission-report_270913.pdf
2.
Kasegn Tekletsadik, Paul Murphy, Scott

Nickerson, Charles Stanley and Semaan Fersan
Progress in Transmission Level Fault Current
Limiters 2013 IEEE PES General Meeting,
Vancouver, BC, Canada, July 21-25, 2013
3.
K. Tekletsadik, P. Lubicki, S. Nickerson, J.

Ludlum, P. Murphy Fault Current Limiter
Selection Considerations CIGRE US National
Committee, 2014 Grid of the Future Symposium,
USA
Fig. 13: SCFCL Recovery Performance Simulation result
SSFCL can be designed to recover instantly, since

there is no component that changes temperature as
quickly during the fault. For applications where the
recovery time has high impact on the system
performance, such as transient stability, voltage
283

April 8-10
2015

SF6 Gas Handling Techniques in Electrical Utilities

Upcoming trends in SF6 gas maintenance and management
Hrushaabh P. Mishra
Syselec Technologies Pvt. Ltd., India
ABSTRACT
SF6 gas has been classified as a potent Green
House Gas by environmental and climate change
organizations. Releasing of SF6 in atmosphere
will be aggravating the current scenario of global
warming and climate change. However, SF6 is
also a popular dielectric used in High Voltage
switchgears, and contrary to the popular notion,
SF6 does decompose and lose its dielectric
strength. This requires quality checks and
regeneration of SF6 in a timely manner.
Commonly used Terms: SF6, Green house gas, Partial
Discharge, PD, Switching, Arcs, Breakdown,
Analysis, Reclaimation.
1.
INTRODUCTION
SF6 gas is an essential gas for electrical equipments

such as circuit breakers, transformers, CTs etc.
Sulphur Hexa-Fluoride or SF6 is a non-toxic, inert,
insulating and a cooling gas of high dielectric strength
and thermal stability. It is particularly suited for the
following applications:
z
design and low noise levels are decisive

advantages.
z
Particle accelerators, X-ray equipments, and

UHF transmission.
2.
CHEMICAL PROPERTIES OF SF6
SF6 is a very stable and inert gas, colourless and

odourless, non-toxic and non-flammable and
insoluble in water. It is one of the least reactive of
all known gases and in normal conditions, it attacks
no known substance with which it will come in
contact
3.
SF6 is 5 times heavier than air and is one of the

heaviest known man-made gases. Due to leaking of
SF6 in air, it will mix insufficiently with air, and will
retain itself near-to-ground levels. SF6 gas cannot
be separated easily from by air. SF6 exhibits 2-5
times better heat transfer properties than air.
Table 1: Physical Properties of SF6
In both High Voltage and Medium Voltage

Power Circuit Breakers: SF6s excellent arc
quenching capacity is put to use here.
In High Voltage Cable: SF6 gas insulated
transmission cables can carry higher capacity
of power
Power Transformers: On account of their high
operational safety, SF6 gas filled transformers
are employed in hazardous operational areas
such as mining. Their light weight, compact
hm@syselec.net
PHYSICAL PROPERTIES
Density
6.14 kg/m3
Thermal Conductivity
0.0136 W/mk
Critical Temperature
45.55 Deg. Cel
Critical Density
730 kg/m3
Critical Pressure
3.78 MPa Absolute
4.
ELECTRICAL PROPERTIES OF SF6
Electronegativity is an attribute of a molecule to

attract and bind electrons towards itself and therb
not allowing the buildup of electron avalanche. SF6
is highly electronegative. It has a pronounced
tendency to bind free electrons forming heavy ions
284

April 8-10
2015

with low mobility making the development of electro

avalanches very difficult.Corona, resulting due to
high potential electric field present in SF6 gas, does
not lead to increase in conductivity within SF6 gas.
Additionally, Sf6 gas is an excellent recombination
gas. It has been observed that SF6 breaks down into
free sulphur and fluorine atoms from 150 Deg. Celsius
onwards. However, as soon as the ambient
temperature is attained, the free atoms of sulphur
and fluorine once again recombine. During arcing of
contacts, the released energy is utilized for
dissociating the SF6 gas molecules, without letting
to further build temperature inside the compartment.
5.
SF6 AS A GREENHOUSE GAS
Coastal territories are to be exposed to

increasing risks due to sea level rise , leading to
increased flooding
A. ISO 14064
In line with IPCC directive and Kyoto Protocol, The
ISO 14064 standard, published in 2006, revolves
around environmental management and quantifying,
monitoring, reporting and verifying greenhouse gas
emissions. The standard is published in three parts:
ISO 14064-1: 2006 specifies principles and
requirements at the organisation level for
quantification and reporting of greenhouse gas
emissions and removals. It includes
requirements for the design, development,
management, reporting and verification of an
organisations greenhouse gas inventory.
ISO 14064-2:2006 specifies principles and

requirements and provides guidance at the
project level for quantification, monitoring and
reporting of activities intended to cause GHGs
emission reductions or removal enhancements.
It includes requirements for planning a GHG
project, identifying and selecting GHG sources,
sinks and reservoirs relevant to the project and
baseline scenarios, monitoring, quantifying,
documenting and reporting GHG project
performance and managing data quality.
Unprecedented combinations of climate change

along with associated disturbances such as
floodings, drought, wildfires, insects, ocean
acidification.
ISO 14064-3:2006 specifies principles and

requirements and provides guidance for those
conducting and managing the validation or
verification of GHG assertions. It can be applied
to organisational or GHG project quantification
including GHG quantification, monitoring and
reporting carried out in accordance with ISO
14064-1 and ISO 14064-2.
6.
ORIGIN OF CONTAMINATION
Approximately 20-30% of plant and animal

species assessed are likely to be at increased
risk of extinction if increase in global average
temperature exceeds 1.5 to 2.5 Deg Cel.
Mishandling during filling and recovering of SF6 are

the major causes of contamination of SF6 with
moisture and air. The residual air present inpipework
and valves, air leaking in through sealings, and by
Stratospheric Ozone Depletion
Average Global Temperature Increase (GHG)
Due to the reflectivity of CO2, H2O, O3, by default

present in the atmosphere, and the increasing
presence of man-made gases such as CO2, N2O
(intensive agriculture), and CFCs(spray propellants
and refrigerants), the IR energy being irradiated from
the earths surface, fails to escape to the atmosphere
and it eventually leads to increase in earths
temperature. With a GWP of 23,900 times that of
CO2, SF6 has been classified as a potent GHG by
Inter-governmental Panel on Climate Change. The
proportion of SF6 has been increasing at an annual
rate of 7%, mostly because of its use in electric
utilities. Once released to atmosphere, it has an
extremely large lifetime of 800-3200 years. In
Europe, SF6 gas falls under F-Gas directive which
bans its usage for all applications except switchgears.
The consequences of Global Warming are far reaching
and some of them have been listed as below:
In dry and tropical regions, like India, crop

productivity is projected to decrease for eben
small local temperature increase (1-2 Deg Cel.)
which would increase the risk of hunger.
Man made pollutants, due to extensive human

activity, have been harming the atmosphere since
the industrial revolution. Two prominent outcomes
of these pollutants are:
A. Contamination during Handling
285

April 8-10
2015

the residual air left inside the enclosure after

evacuation and before filling with SF6 go unnoticed
and eventually dilute the purity of SF6. The ingress
of air and gas-entrained dust can be minimised by:
z
Appropriate Designwork of pipes and valves
Appropriate handling procedures
Careful evacuation of the air from the enclosure

before filling with SF6 (a residual air pressure
of 1 millibar is recommended)
B.
Contamination by Leakage
The diffusion of air and humidity might happen from

the outside to within the compartment because the
partial pressure of air and water outside the enclosure
is higher than that inside. The main leakage paths
are enclosure porosity, sealings of mechanically
moving transmission elements, and O-ring sealings.
SF6 half-empty cylinders over a period of time may
become corroded and might start leaking. The
pressure within them may become destabilized,
which might cause a potential burst too.
C. Contamination by Desorption
During the assembly of the equipment itself,
accumulation of Humidity and gases within the inner
surfaces of the equipment or in the bulk materials
are released within SF6 during its normal operation
and at elevated internal temperature. Polymeric
materials such as epoxy are highly hygroscopic and
have poor moisture retention capability at increased
temperatures. Adsorbers which have not been
properly handled may contain both humidity and
adsorbed SF6 decomposition products which can be
released during evacuation or at elevated
temperature. The quantities of the desorbed
substances are difficult to estimate because they
depend on the specific materials employed,
production methods, quality control and the
assembling and maintenance procedures.
D. Decomposition by Electrical Discharges
Partial Discharges of the corona type
Spark Discharges
Switching arc
Failure Arcs
Partial Discharge
PD is defined as a localised breakdown. PD activity

is initiated within SF6 gas due to the presence of
moisture and air, protrusions or abrasions on the
insulator surface, freely floating particles within the
gas. Corona is defined as a type of Partial Discharge
that occurs in gaseous medium around the conductor.
It is the ionisation of the surrounding gas and usually
develops around the freely floating electrodes.
Corona is an electron dominant process. The electron
temperature is much higher than the gas temperature,
since the gas exists under non-equilibrium
conditions. The mean energy of the electrons in the
corona is limited to 5 to 10 eV. This already exceeds
the SF6 bond energy between SF5 F of 3.5 4 eV
and the electron impact will dominate the
decomposition process. The electron impact
dissociation process can lead to the following
reactions:
e + SF6 > SFX + (6-x)F, x <5
e + SFX > SFx + F
Multi-step dissociation leads to the formation of
lower fluorides of Sulphur such as SF4 and SF3. In
the absence of contaminants in the gas or on the
surface of insulators, the products of decomposition
of SF6 quickly recombine though relatively rapid
process to explain the thermal conversion of SF4 +
F2 into SF6. However, in the process of oxygen and
moisture, the recombination process can be
interfered with by the reaction between the lower
fluorides of sulphur and the contaminants to form
sulphur oxyfluoride, HF and metallic fluoride. Some
of the reactions can be expressed as follows:
SF5 + OH > SOF4 + HF
SF5 + O >SOF4 + F
SF4 + O > SOF2 + 2F
SF2 + O > SOF2
SF4 + OH > SOF2 + HF + F
SF6 is partially decomposed by electrical discharges

which can be grouped into four major types namely:
z
E.
SF3 + O2 > SO2F2 + HF

SF4 + H2O > SOF2 + 2 HF
SOF2 + H2O > SO2 + 2 HF
SOF4 + H2O > SO2F2 + 2 HF
Outside the corona discharge region, slower reactions
286

April 8-10
2015

between the long-lived lower sulphur fluorides such

as SF4 and stable oxyfluoride with the contaminants
will further lead to the formation of other types of
compounds such as SO2, SO2F2, and SOF2.
is due to eroded PTFE (a CF2 polymer), which is

employed in most switchgear to contain the arc.
H. Failure Arcs
Failure Arcs are a result of insulation breakdown or
switch gear interruption failure and occur extremely
rarely. In these events, the arc burns mainly between
metallic minerals, which are not designed for arcing
such as aluminium, copper and steel. These materials
have relatively high arc erosion rates.
I.
Fig. 1: Traces of flashover on a busbar (left)and surface carbon
deposition due to failure arcs (right)
F.
Spark Discharges
Spark Discharges may occur at large scale insulation

defects such as floating conductors and during
disconnector switching operation. The
decomposition products generated are of the same
kind as in corona discharges but their quantitative
generation rates and compositions are different.
G. Switching Arcs
Switching arcs occur during the load
breakingoperation of power circuit breakers. At the
centre of high current breaking arc, the temperature
rises to be as high as 20000 K. As the arc cools,
recombination of sulphur and fluorine to form SF6
occurs rapidly, often in the microseconds scale. In
the presence of oxygen, H2O and metal vapour,
resulting of electrode heating, the recombination
process is altered which leads to the formation of
various arc by-products. The high current flow in
these arcs leads to substantial erosion of the contacts
and insulation material by the hot arc. The main cause
for SF6 decomposition is the reaction of these erosion
products with the fragments of thermally dissociated
SF6 and other trace gases such as moisture and
oxygen. The most important of these reactions can
be expressed by the below three formulae:
Mechanical Generation of Dust Particles
Metal dust particles may be generated by mechanical

friction of metal surfaces. In properly designed
equipment, these particles usually fall into areas
where they have no effect on the insulation integrity
of the installation. If however they fall into the area
of high electric field stress such as an insulating
barrier, they may cause tracking on the insulator
surface and flashover.
7.
SF6 IMPURITIES
The impurities resulting in the SF6 gas are Air,

Moisture, HF, CF4, and SO2. Permissible limits have
been laid for the aforementioned impurities according
to IEC60376 and IEC60480. IEC60376 lays down
the guidelines for checking brand new SF6 gas and
IEC60480 lays down the guidelines for checking used
SF6 gas.
Cu + SF6 > CuF2 + SF4

W + 3 SF6 > WF6 + 3 SF4
CF2 + SF6 > CF4 + SF4
The first two reactions are associated with eroded
material from the arcing contacts for which copper
tungsten (Cu-W) is normally used. The last reaction
287
Table 2: Maximum Acceptable Impurity Levels

for new SF6 Gas (or gas filled in cylinders)IEC 60376
Impurity
Specification
according to 60376
Air
0.05% by volume
CF4
0.05% by volume
H2O
122 ppmv
Table 3: Maximum Acceptable Impurity Levels for

Reuse of SF6 Gas - IEC60480
Impurity
Specification
according to 60480
Air and /or CF4
3% by volume
H2O
200 ppmv
HF
25 ppmv
SO2
50 ppmv

April 8-10
2015

impact. However, the dielectric strength of a mixture

of SF6 with Air or with CF4 is always lesser leaser
than that of pure SF6. SF6 mixture with air/ CF4 has
greater arcing time and the interruption derating
occurs. If the performance has to be maintained, then
the pressure will have to be risen or the redesign of
the arc extinguishing zone would need to be done.
D. Heat Transfer
z
Fig. 2: HF dust accumulated at the bottom of HV contact plate
8.
EFFECTS OF CONTAMINATION
A functional deterioration of equipment by SF6

contaminants can, in a general sense, be viewed in
the following six respects:
z
Health risk
Corrosion
Insulation Performance of gas gaps
Insulation Performance of insulator surfaces
Switching capability (for switchgear only)
Heat transfer
A. Health Risk
SO2 and HF constitute a health risk, whereby they
might cause irritation effect in eyes and nose and
lead to difficulty in breathing. HF, one of the highly
corrosive gases, causes severe skin burns if it comes
in contact with the skin of human personnel.
B.
Corrosion
HF, corroding the contacts and the internal metallic

parts of a gas compartment will lead to the formation
of metallic fluorides in form of dust particles.
Reaction of SO2 and HF with insulator surfaces leads
to pitting, and decrease in surface resistivity. This
results in surface tracking and increased
conductivity.
Air and CF4
Modern Designs however avoid the use of corrosion

sensitive materials. Some of the contaminants are
chemically inert such as Air, CF4 and moisture may
affect the gas insulation capability and the circuit
breaker switching performance, if present in too high
concentrations. They may also have an influence on
the convective heat transfer by the insulating gas.The
main conducting liquid contaminant is water
condensing from moisture in the form of water
droplets or films. As water has an extremely high
dielectric constant and high electric conductivity, it
causes local field enhancements at droplets and
conducting surface films along insulators both of
which deteriorate insulation performance . Moisture
is mainly introduced by desorption from surfaces
and from the bulk of polymers. Its condensation is
controlled by the absolute moisture which is
expressed as partial water vapour pressure
pH2O.Non-conducting solid decomposition products
are generated from arc eroded metals by reaction
with dissociated SF6. They mainly consist of copper
fluoride CuF2, tungsten oxide WO3, the tungsten
oxyfluoridesWO2F2 and WOF4 originating from
switchgear contact erosion and Aluminium Fluoride
AlF3 in case of internal arcing. They are non-critical
for insulation as long as they are not exposed to
excessive moisture. Conducting solid contaminants
such as carbon and metal dust may become critical
when deposited on field exposed insulator surfaces
as conducing layer. Carbon may be generated by
carbonisation of polymeric materials. Metallic dust
particles generated by mechanical friction may be
transported by gas flow.
C. Insulation Performance
9.
Air and CF4 are chemically inert gases. SF6 is usually

diluted with these two gases in order to reduce the
release of SF6 and the associated environmental
Infra-red based analyzers require the smallest amount

of gas for analysis. It gives the readings at the fastest
288
INFRA-RED TECHNIQUE BASED SF6

GAS ANALYSIS

April 8-10
2015

rate with no cross interference or contamination.

Infrared based SF6 gas analyzers are preferred today
since they differ from the conventional sensors on
following grounds:
Table 4: Comparison between Conventional and
Ndir Technique
Conventional
Technique
based analysis
NDIR (Non-dispersive
Infrared)
Spectroscopic Analysis
Less reliable as the

electrolyte has to come
in contact with the
analyzed gas
Highly reliable as it is a
non-contact based
measurement
Highly susceptible to
temperature and
humidity drifts
Temperature and humidity

compensation is
possible
Less accurate due to

invariable response.
This leads to false
alarms.
Highly accurate due to

minimal drifts due to
temperature & humidity
effects. This leads to
accurate alarms
Slow responding
Very fast in responding

and producing results
HF detection is not
possible as it corrodes
the sensor
HF detection is possible
since it is a non-contact
based measuring method
Affordable cost wise
Expensive technology
the gas cannot be purified sufficiently on-site to

achieve the reclaimed criteria it still has reuse
potential when treated by a specialised processing
company. This allows it to be rendered reusable in
the majority of cases and some SF6 producers are
already offering such a purification service. Only a
very small fraction of the residue will have to be
processed for the final disposal in an environmentally
compatible way.
A. Reclaiming Equipment
Gas reclaimers have been used successfully since the
introduction of SF6 technology. They are
commercially available in a variety of sizes, gas
processing capacities and storage capacities and range
from units that can be hand carried to larger trailer
mounted systems. The appropriate type and size of
the reclaimer should be chosen according to the gas
capacity to be handled.
B.
Block diagram of a Reclaimer
The basic functional schemes of a reclaimer are as

follows:
10. REMOVAL OF THE SF6 IMPURITIES

While SF6 itself is not consumed in significant
quantities, its performance can degrade due to
contamination by air, moisture, decomposition
products. It is desirable both from an ecological and
economic point of view, to keep SF6 at a low
contamination level by careful handling so that it
can be reclaimed on site many times. SF6 continually
should be reused during equipment development,
product testing, commissioning, maintenance and
repair and decommissioning where the criteria of
IEC60480 can be achieved. It thus goes through a
continuous cycle of reuse. Such a systematic re-use
of SF6 requires that the gas be kept at the stated
quality level at which it can perform its functions.
With properly maintained reclaiming equipment,
humidity and reactive decomposition products can
almost always be removed on-site so that nonreusable gas can be transported. In the rare case that
C.
Filters
The various types of filters used in reclaiming carts

are as follows:
289
Filter type
Particle Filter
Tasks
Removes solid
decomposition
products and
other particles
at input of
reclaimer
Major
Characteristics
Pore size 1 m

April 8-10
2015

Gas/Humidity Removes
Filter
Gaseous
Decomposition
Products
Residual Humidity
<100ppmv
Residual SO2
<10 ppmv
Residual Reactive
products
Detoxification Reduces
Filter
reactive
decomposition
products
Same as prefilter
Oil Filter**
Special Filter
equipped with Visual
Oil Indicators
Removes Oil
processed by the reclaimer. It must have enough

capacity to store the amount of SF6 to be recovered.
It can be an integral part of reclaimer or an external
item. If used for liquid storage of SF6, it should be
rated for a pressure of atleast 50 bar.
H. Refilling
The reclaimer must have the provision to allow
refilling of gas, from the storage vessel into an
electrical equipment, the refilling procedure
requirements will vary according to storage method
employed.
I.
**- will not be required if compressors are oil-free
Hose Connection
Adsorbents
Contaminants Removed
Molecular Sieve 4A
Water, SO2, SOF2, SF4
Hose connections should be self sealing to prevent

air and moisture from entering the gasreclaiming
equipment. As the equipment will often be left in a
state of vacuum these SF6 valves needto be pressure
and vacuum tight
Molecular Sieve 13X
Water, SO2, SF2, SF4(SFn)
J.
Activated Alumina
Water, SO2, SOF2, SF4, HF
Soda-Lime
(CaO-NaOH)
Water, SO2F2,HF
Activated Charcoal
Oil Vapour
Gas piping and pipe unions used must be of high

quality and preferably use a metal to metal re-usable
sealing system, proven in its performance with SF6
and its decomposition by products. The following
characteristics are essential:-
Table 5: Typical Adsorbents for Various SF6

Impurities
Pipe Work
D. Vacuum Pump
pressure and vacuum tight
The vacuum pump module is used to extract air from

SF6 insulated equipment and associated piping prior
to refilling with SF6 and for dehydration (removal
of residual moisture). It is also used to remove air
from various sections of the gas processing system
itself, eg, after maintenance work.
vibration proof
re-usable (indefinite refitting possible)
temperature change resistant
E.
SF6 Vacuum Compressor
The SF6 vacuum compressor module is used to

recover SF6 from gas insulated equipment and to
assist the series connected SF6 piston compressor.
F.
Gas/Humidity Filter
Filter elements should be adequately sized to remove

moisture, gas by-products and particles larger than
1 micron in size (as a second means of trapping and
larger particles which may have been transmitted
via a non efficient particle filter).
G. Storage Module
The SF6 storage module is used to store SF6
All pipework should be of copper tubing silver

soldered to tube unions and having inline
covering of PTFE (Polytetrafluoroethylene).
All components (gauges,valves, filters, etc.) should
be securely mounted to the frame of the gas cart,
such that pipework doesnot have to support
them. This prevents stress cracks causing either gas
losses or inadvertent gasmixtures. Heavy
components (i.e. compressors and vacuum pumps)
should be shock-proof mountedand be connected
to the fixed pipework via flexible connections.
Equipment used to process SF6 (i.e. compressors)
should be oil less, of gas tight construction andshould
not contain any internal components that can be
corroded by decomposition products (e.g.galvanised
metal).
290

April 8-10
2015

11. CONCLUSION
3.
SF6 gas can be reused for electricity utility field

purposes by periodic checking of the quality of the
gas with a NDIR based instrument. If impurities are
found to be beyond permissible limts, onehas to do
purification of gas for reusing it.
IEC 1634, Technical Report, Use and Handling

of SF6 in High Voltage Switchgear and Control
Gear, First Edition, 1995-2004.
4.
IEC 60480, Guide to the checking and

treatment of SF6 takn from Electrical
Equipment.
5.
Solvay Fluor, Sulphur Hexafluoride Brochure.
REFERENCES
1.
2.
P Glaubitz (Convenor), S Stangherlin

(Secretary), D Crawley, J Henriot, A Holm, P
Jannick, P Justiz, Probst, P Sieber, S Theoleyre,
T Yokota, L Van der Zel, Guide for the
preparation of customised Practical SF6
Handling Instructions, CIGRE No. 276, TF
B3.02.01, August 2005 .
G Mauthe (Chairman), B M Pryor (Secretary), L
Niemeyer, R Probst, J. Poblotzki, H D Morrison,
P Bolin,P OConnell, J Henriot, SF6 recycling
guide, CIGRE 23.10 TF 01, August 1997.

AUTHOR
Hrushaabh Mishra is Technology Resource Head
with Syselec Technologies Pvt. Ltd. He is a Gold
Medallist in B.Tech. Electrical Engineering from
Government College of Engineering, Amravati, M.S.
He has been specializing in SF6 Gas Testing services
for reputed utilities the past 4 years. His other areas
of interests include Partial Discharge testing for GIS,
Power Transformer Residual Life Assessment and
Switchgear protection testing.
291

April 8-10
2015

Advanced Method for on Line Monitoring of Power Transformers for

, Transients & Partial Discharge
Bushing C / Tan
G.S. Papneja
Sridhar Shenoy
OMICRON Energy Solutions Pvt. Ltd, India
Wojceich Koltunowich
OMICRON Energy Solutions, Germany
SUMMARY
Failure statistics of power transformers show that
the majority of failures are caused by insulation
defects. Almost half of these insulation problems
originate at the bushings. A commonly used DGA
(dissolved gas analysis) sensor detects insulation
problems inside the tank only, but not at the
bushings. Therefore supplementary online bushing
monitoring is vital.
This paper describes a new on-line method for
capacitance and dissipation factor measurements
of high voltage bushings, for recording grid
transients and for partial discharge monitoring at
power transformers. The needed accuracy for
reliable detection of failures is in the range of 5
pF for capacitance and 0.1 % for dissipation factor.
For reaching this goal, a relative comparison of
the bushing currents of the three phases, measured
at the bushing taps, does not provide sufficient
accuracy due to imbalances in the network.
Additionally, simultaneous ageing, as it can be
anticipated for a three phase transformer, will not
be detected. To overcome these shortcomings, this
paper introduces an absolute method for
measuring capacitance and dissipation factor of
HV bushings. Even under field conditions
measurement accuracy similar to accurate off-line
power factor tests was reached. Partial discharges
are an early breakdown indicator for HV
dielectrics. With the measurement on the HV
bushing tap, not only failure of the bushings
dielectric but also of the transformer insulation
system can be detected. As for off-line
measurements, the discrimination between noise
and true internal discharges is the key for
successful diagnosis. Strategies for sensitive
partial discharge measurements under disturbed
on-line conditions are described, also involving

measurements in the UHF range.
Keywords: power transformers, on-line monitoring,
bushings, dissipation factor, capacitance, partial
dischcarges, UHF
1.
ON-LINE MONITORING OF POWER

TRANSFORMERS
For many decades, transformer users have sought

ways to assess the general condition and identify
specific problems of their assets. Conclusively,
diagnostic tests have been developed, which are
applied in the de-energized status. In the last years,
a sophisticated means has evolved for collecting a
great deal of diagnostic information while the
equipment is in service. While periodic off-line
diagnostic tests still play the dominating role in
condition assessment, continuous or on- line
monitoring promise to have the potential to
overcome some of the fundamental limitations of
off-line tests:
z
Continuous measurement for having reliable

measurement data, reducing the effect of
outliers and for continuous observation of the
equipments condition;
Early diagnosis of initiating failures for
scheduling maintenance actions, therewith
supporting condition based maintenance
schemes;
Knowledge of the equipments historical use for
fully utilizing the life span in the context of asset
life management.
Beside these advantages, on-line monitoring of

power transformers is apparently becoming an
essential feature of the smart electric utility
292

April 8-10
2015

systems of the future.

Still, not all of these promises are fulfilled with
todays monitoring solutions, as experts discuss
questions like How to get real value from
continuous on-line monitoring systems? Error!
Reference source not found..
This paper describes a new method for monitoring
of capacitance and dissipation factor (power factor)
of transformer bushings, transients in the HV grid
and partial discharges in the transformer and its
bushings.
Particular focus is given on an extraordinary high
accuracy of all measurement values; an accuracy,
which allows to move forward from a warning
system to advanced diagnostics; in some functions
even to replace off-line tests.
essentially with about 20 %. Failures in the winding

are often emitting partial discharges and can be a
result of transient over-voltages in the electrical grid.
Thus, a monitoring system for the bushings
capacitance and dissipation factor, partial discharges
and grid transients will likely lower the failure
probability of the complete asset.
1.1 Bushing Failure Modes and Requested

Sensitivity
A monitoring system for HV bushings should early
detect common failure mechanisms for transformer
bushings.
Partial breakdowns between field grading layers result
in an increase of capacitance and in partial
discharges. Table 1 gives the increase of capacitance
for bushings of different voltage levels with their
typical number of field grading layers.
Table 1: Voltage class and change of capacitance
for condenser type bushings
Since bushings are subjected to high dielectric and

thermal stresses; bushing failures are one of the root
causes of forced outages and transformer failures.
Figure 1 shows transformer failure statistics,
compiled by the authors of Error! Reference source
not found. from 7 other sources. Bushings contribute
Voltage in kV
No. of layers
Change in %
123
14
7.1
245
30
3.3
420
40
2.5
550
55
1.8
Ageing and moisture increase the dissipation factor

DF or power factor PF. Standards as in Table 2 give
advice on evaluating the test results for resignimpregnated-paper RIP, oil-impregnated-paper OIP
and resign-bounded-paper bushings. Consequently,
monitoring systems should be able to provide an
appropriate accuracy to evaluate the bushings
condition according to the relevant standards.
Table 2: Acceptance level of dielectric losses
for bushings of different design
Standards
RIP
OIP
RBP
DF IEC60137
<0.7%
PF IEEE C57.19.01
<0.85% < 0.5 % < 2 %
<0.7% <1.5%
Further failure mechanisms are voids and cracks,

resulting in partial discharges and a slight change of
capacitance and finally corroded contacts, resulting
in partial discharges.
Fig. 1: Reasons for transformer failures
293

April 8-10
2015

From the considerations above we can draw the

conclusion that a monitoring system with sensitive
measurement of capacitance, dissipation factor and
partial discharges has the capability of an early
diagnosis of all typical initiating failure modes.
2.
MONITORING SYSTEM DESIGN
2.1 Accurate Measurement of Capacitance

and Dissipation/Power Factor
Traditionally, the sum-of-currents method is the most
common method used in monitoring of bushing
capacitance. The leakage currents measured at the
three bushing taps are added to obtain the imbalance
current. The changes of the system voltage phase
angle due to the daily load variation lead to
significant changes of the bushing capacitance and
dissipation factor, introducing a measuring error 1.
To increase the sensitivity of the measurements,
reference and comparison methods have become
more and more popular 2. Both these methods require
a reference signal for calculation of dissipation factor
and capacitance. Depending on the source of the
reference signal, relative or absolute measurements
can be performed (Figure 2) 3 . With relative
measurements, the bushing-to-bushing comparison
or dual transformer comparison is performed, while
with absolute measurement reference signal is taken
from voltage transformers (VTs) of the substation.

Grid unbalance, like phase shifts or changes in the
system voltages can have a misleading impact on
the relative measurements of capacitance and
dissipation factor. The simultaneous ageing of the
reference and test bushing cannot be detected. To
reduce the impact of these uncertainties (Figure 3),
absolute measurements are recommended at the
Winchelsea substation. The reference signal is taken
from a group of VTs located near the monitored
transformer. This pure resistive signal (UVT) is
compared with the mainly capacitive leakage current
(IB) measured at the bushing tap of the transformer.
The corresponding vectors of these two signals are
rotating in the same four-quadrant coordinate
system
r
(Figure 4) and the capacitive current I B always leads
in revolution. The angles between the reference axis
and the two vectors are permanently measured and,
by computing the difference between them, the
angle and the power factor cos are obtained. The
dissipation angle () and dissipation factor (tg) can
be calculated accordingly. The temperature
correction of the dissipation factor values is
performed [4]. It takes into account the OIP (oil
impregnated paper) type of the bushing. The formula
(1)
is used to calculate the bushing capacitance CB.
The IB current is measured by the monitoring system,
while the
value is directly taken from VTs.
Fig. 2: Architecture of the monitoring system applied for comparison and reference method
294

April 8-10
2015

Table 3: Voltage error and phase displacement for

measuring voltage transformers 5
Percentage
voltage (ratio)
error
Minutes
Centiradians
0.1
0.1
0.15
0.2
0.2
10
0.3
0.5
0.5
20
0.6
1.0
1.0
40
1.2
3.0
3.0
Not
Specified
Not
Specified
Class
Phase Displacement
2.2 Combination of Conventional and

Unconventional PD Detection
Fig. 3: Sensitivity/uncertainty introduced by each method
Fig. 4: Vector representation of the UVT and IB in a four-quadrant

coordinate system
I B = 2 f C B U High C B =
IB
2 f U High
....(1)
The accuracy of the monitoring system is < 2 pF

plus dyn. VT error for capacitance measurement and
0.05 % plus dyn. VT error for dissipation factor
measurement. The class of VTs also has to be taken
into consideration (Table 3).
Two methods are recommended for PD

measurements: the conventional method (according
to IEC 60270) with sensors at the bushing taps and
an unconventional ultra-high frequency (UHF)
method with a sensor placed inside the transformer
tank. With the conventional method, the PD signal
from each tap is synchronously acquired by a three
channel acquisition unit. The central frequency of
the digital band pass filter of the acquisition unit is
selected to reach the optimal signal-to-noise ratio.
To obtain more detailed information about the type
and location of the insulation PD defects, the
unconventional UHF PD measurements in the
frequency range between 100 MHz and 2 GHz with
an antenna installed inside the oil drain valve are
performed. The presence of external noise in this
frequency range is low and radio or mobile phone
signals are easily recognized and eliminated from the
measurements. PD activity inside the bushings
insulation and close to the end winding area is mostly
detected with the conventional method while the
rest of the tank is covered by the UHF antenna. Even
if the PD signal is measured in a different frequency
range, the obtained PRPD diagrams have very similar
patterns that makes recognition of the defects type
easier (Figure 5). The signal detected by the UHF
antenna is synchronized with the signal detected at
the bushing taps. Furthermore, the measured pulses
in the UHF range, mostly coming from internal PD
activity, can trigger the start of conventional
measurements. Thus a better separation between
internal and external PD pulses can be obtained.
295

April 8-10
2015

Fig. 5: Example of PRPD patterns measured by conventional method (left) and UHF method (right) for a known internal PD defect
2.3 Measurements of Transient Over-Voltages

The voltage transients appearing at the bushings and
propagating into the windings create thermoelectrical stress and accelerate the aging of the
insulation. The number of the transients and their
shape (rise time, magnitude and duration) are
important parameters. The magnitude of the
lightning-induced transients is reduced below 2 p.u.
(phase-to-ground voltage = 1 p.u.) by surge
protective devices 6,7, but the fast transients coming
as a result of operations in the substation can also
be dangerous 8. The magnitude and shape of the
transients strongly depends on the power grid
configuration, length of the lines, type of connecting
HV equipment 6. Figure 6 shows the waveform of
the transients when a cable, gas insulated line (GIL)
or overhead line (OHTL), all with the same length of
50 m, are connected at the transformer. In Figure 7,
Fig. 7: Variation of peak magnitude of transients as a function of

the length of terminals 7
the variation of transient peak magnitude as a

function of the length of connecting terminals are
presented.
3.
ANALYSIS OF THE RESULTS
3.1 Capacitance and Dissipation Factor

The example of the trend of capacitance is presented
in Figure 8. The maximum value of C measured
over the period of several months is compared with
the threshold value (Figure 9). The C in phase B is
the highest but still below the maximum 10 % limit
of change.
Fig. 6: Waveform of the transients when a cable or GIL or OHTL

is connected at the transformer 7
The three-phase trend of the dissipation factor is

shown in Figure 10. All the values are below the
threshold values predefined for OIP bushings 9 and
their trend is stable. A comparison between the
296

April 8-10
2015

highest on-line measured value for each phase and

the predefined threshold is shown in Figure 11.
4.
PD MONITORING
4.1 PD Monitoring with the Conventional

Method
The three-phase PD trend is presented in Figure 12.
At each point visible in the trend diagram, PRPD
patterns and 3PARD are available.
Fig. 8: Three phase capacitance and ambient humidity variation

(red trend)
Fig. 9: Comparison between the highest capacitance variation

(green) and the threshold limit (red)
Figure 13 depicts the PRPD patterns of the PD signal

acquired by the three-channel synchronous system.
They are complex patterns with several PD sources
overlapped. In order to separate clusters of different
PD sources, a synchronous multi-channel PD
evaluation technique is applied [11]. The 3PARD
diagram (Figure 14) visualizes the relationship
between amplitudes of a single PD pulse in one phase
and its crosstalk generated signals in the other two
phases. By repetition of this procedure for a large
number of PD pulses, PD sources within the test
object as well as outer noise appear as a clearly
distinguishable concentration of dots in a 3PARD
diagram [11]. By examining individual clusters in the
3PARD diagram, a separation between noise and PD
phenomena is possible.
Figure 15 shows the back transformation to PRPD
pattern of the clusters 1 and 2. The patterns of the
clusters 1 and 3 appear to be generated by bubbles
and surface discharge with the highest amplitude in
phase B (cluster 1) and phase A (cluster 3). Similar
PRPD patterns were reported in [12]. The shape and
phase position of the patterns of the clusters 2 (phase
A) and 6 (phase C) may indicate partial discharge
activity inside the voids of the insulation system [12].
The clusters 4 and 5 are generated by external
interferences
Fig. 10: Three phase dissipation factor variation
Fig. 11: Comparison of the highest measured value of dissipation

factor with the threshold limit
Fig. 12: Three phase PD trend
297

April 8-10
2015

Fig. 13: Three phase synchronous PRPD patterns
4.2 PD Monitoring with Unconventional UHF Method

Performing a frequency sweep, two spectra of the signal are Obtained (Figure 16). The upper spectrum is
built based on the maximum amplitudes of the time domain signal acquired at each value of the frequency
during the sweep. The lower spectrum corresponds to their minimum amplitudes. Internal PD activity is
always visible on the upper spectrum while external interferences (corona discharge, radio waves, GSM) are
visible on both spectra.
Fig. 14: Equivalent 3PARD diagram

Fig. 15: Individual PRPD patterns of the selected clusters
Internal PD activity was identified in the frequency

range from 450 to 650 MHz. The PRPD pattern
corresponding to a central frequency of 600 MHz is
presented in Figure 16. The signal detected during
the on-line monitoring was synchronized with a 50
Hz signal taken from the measuring tap of the phase
A. It can be seen in Figure 17 that the phase of the
voltage where the PDs occur is the one characteristic
to internal discharges. Furthermore, it indicates a
possible location of the PD activity, namely, in the
vicinity of phase A. The PRPD patterns of frequencies
between 1 GHz and 1.4 GHz were checked and no
internal PD activity was found.
298
Fig. 16: Frequency sweep diagram

April 8-10
2015

4.3 Transients
Figure 18 shows an example of transient overvoltages with the amplitudes between 1.25 and 1.77
p.u. recorded in four months after installation of a
monitoring system.
The transient with the highest amplitude (1.77 p.u.)
is shown in Figure 18(a), Figure 18(b) and 18(c) show
other transient over-voltages recorded during the
monitoring period. Even with such amplitudes, these
transients represent a threat to the insulation of the
bushings and windings because of the resonant
phenomena which can lead to higher voltage
distribution between the turns. The number of
transients cannot be controlled but their monitoring
can help identify defective equipment which
generates over-voltages, near the transformer.
Fig. 17: PRPD pattern
Fig. 18: Time stamp and shape of the transients detected at the bushings
5.
CONCLUSIONS
REFERENCES
This paper describes a new on-line monitoring

system for measuring bushing capacitance and
dissipation factor, grid transients and partial
discharges of transformer and bushing. While using
voltage transformers as reference, it provides an
accuracy comparable to off-line bridge instruments.
For partial discharge measurements, various methods
for discriminating between internal discharges and
external noise are described: UHF gating and
correlation diagrams based on amplitude ratios and
frequency content. The on-line monitoring system
aims to provide reliable, automated on-line
monitoring functions for observation and warning
and, with the same hardware, an advanced diagnostic
tool in case of insulation failures.
1.
Z. Berler, V. Prykhodko, J. Watson, J.S. Skinner,

D. Bates, Analyzing Data from On-line
Continuous Bushing Monitoring Systems Paper
9, proceedings of the International Conference
on Condition Monitoring, Diagnosis and
Maintenance CMDM2013, Bucharest,
Romania, pp.73-78, 2013.
2.
M. Koch, M. Krger and W. Koltunowicz,A New

Method for On-line Monitoring of Bushings,
Transients and PD of Power Transformers,
Proceedings of The 4th International Advanced
Research Workshop on transformers, Baiona,
Spain, pp. 287-290, 2013
3.
Transformer Bushings, type GOB Installation
299

April 8-10
2015

and Maintenance Guide, ABB Technical

Brochure,http://www.abb.de/productguide/
product.aspx?&c=c12573e700330462c1256f5
c004a7dc9&db=db0003db004283
4.
***IEC60044-2:2003 Instrument transformers

Part2: Inductive voltage transformers
5.
X. Dong, S. Rosado, Y. Liu, N.C. Wang, E.L. Line

and T.Y. Guo, Study of abnormal electrical
phenomena effects on GSU Transformers, IEEE
Transactions on Power Delivery, vol. 18, No. 3,
July 2003.
6.
A. Said, E.A. Badran, M.A. Abd-Allah,

Mitigation of very fast transient overvoltages
at the more sensitive points of gas insulated
substation, International Journal on Electrical
Engineering and Informatics, Vol. 4, No. 3,
October 2012.
7.
U. Riechert, H. Ito, E. Zaima, K. Uehara and W.

Chen, Insulation Co-ordination for Very Fast
Transients in Gas-Insulated UHV Substations,
presented at UHV Colloquium New Delhi 2013,

Session 2.3 Substations.
8.
CIGRE TB 445 Guide for Transformer

Maintenance CIGRE Working Group A2.34,
February 2011.
9.
T. Stirl, R. Skrzypek, S. Tenbohlen, R.

Vilaithong, Online Condition Monitoring and
Diagnosis for power Transformers their Bushings,
Tap Changer and Insulation System Paper?
Proceedings of the International Conference on
Condition Monitoring and Diagnosis CMD 2006,
Changwon, South Korea , April 2006.
10. W. Koltunowicz and R. Plath, Synchronous

Multi-Channel PD Measurements IEEE
Transactions on Dielectrics and Electrical
Insulation, Vol.15, No. 6, p. 1715-1723, 2008.
11. A. Carlson, J. Fuhr, G. Schemel, F. Wegscheider,
Testing of Power Transformers Routine tests,
Type tests and Special tests, 1 st Edition,
published by Pro Print, Zrich, Switzerland,
2003.
300

April 8-10
2015

Feedback on In-Service Deployment of the

Fully Digital Substation
G. Lloyd, S.H. Richards
J. Wright
Alstom Grid, UK
M. Boucherit
A. Procopiou
Alstom Grid, France
SUMMARY
Digital devices such as numerical protection
relays and digital systems such as substation
control are prevalent as the technology of choice
in substations today. This article looks at other
substation areas into which digital technology has
permeated, such that progressively, substations
can become fully digital in their implementation.
Denmark is taken as an example, having a
significant number of windfarms on a number of
islands. These are interconnected using 400kV
hybrid lines. To avoid disfiguring the beauty of
certain environmentally protected areas,
underground
cables
were
installed.
Nonconventional fully-digital instrument
transformers were selected due to their low mass
and slimline construction, replacing conventional
wound current transformers. Correct application
of IEC61850 standards allows a lot more of the
substation engineering to be undertook in the
controlled environment of the factory, rather than
on the substation site where outage times and
site labour are at a premium.
The paper builds on the in-service advantages
realized using IEC 61850-8-1 for the DCS,
including full protection and control, extending
the implementation to include the process bus.
Process bus deployments replace the traditional
hardwiring to primary equipment with an Ethernet
link, and also convert the primary current and
voltage measurement channels to protection
relays and other IEDs (intelligent electronic
devices) to fibre-optic too. Digital implementation
helps to reduce the substation physical size, move
as much configuration and testing to the FAT stage
(factory acceptance testing), and also decouples
Ritesh Bharat
Alstom T&D, India
the dissimilar renewal cycles of the primary and

secondary equipment.
Keywords: Ethernet, Station bus, Process bus, IEC
61850, Merging Unit, Digital Instrument Transformer
1.
INTRODUCTION
Firstly, one might ask what is a digital substation?,

and this will yield a variety of possible replies, as
there is no standard definition. Clearly as most
substations today are switching and routing AC
power at high/extra high voltage, it is not the primary
flow which is digital. This means that we are talking
about the secondary systems, and all the protection,
control, measurement, condition monitoring,
recording and supervisory systems associated with
that primary process.
In general terms, a full digital substation is one in
which as much as possible of the data related to the
primary process is digitised immediately, at the point
where it is measured. Thereafter, the exchange of
that measured data between devices which may need
to subscribe to it is via Ethernet, as opposed to the
many kilometres of copper hardwiring which may
exist in a conventional substation.
Digital substations imply a solution and architecture
in which the substations functionality is
predominantly now achieved in the software, with
lesser reliance on hardware implementations such
as hardwiring.
2.
DRIVERS TOWARDS DIGITAL

SUBSATIONS
Some of the areas in which digital technology brings

real benefits are now highlighted:
301

April 8-10
2015

Increased reliability and availability
The extensive self-diagnosis capability of digital

devices ensures maximised up-time of the substation.
Any degradation in the performance of an asset is
pinpointed in real-time. Inherent redundancy in the
system may be employed to self-heal the operation,
which permits troubleshooting without the need for
any primary system outage.
z
Optimized operation of assets
The intelligence within digital substation schemes

allows close monitoring of the loadflow capacity of
plant equipment, compared to its design ratings.
z
Improved safety
The removal of wired cross-site current transformer

circuits reduces the risk of lethal injury due to
inadvertent opening of the circuit by personnel. The
avoidance of oil in instrument transformers reduces
explosion risks too.
z
Reduced maintenance costs
The digital substation closely monitors all substation

assets. Intelligent systems analyse the data and
provide recommendations on maintenance and repair
actions to conduct. This allows a shift to predictive
or reliability-centered maintenance, avoiding
unplanned outages and emergency repair costs. It is
in this regard that ongoing operational cost savings
may be made, such that the lifetime total cost of
ownership of the substation will be reduced.
z
Easier renovation and extension of existing

substations
Interoperable solutions and the use of fiber optics

instead of copper wires reduce the duration and cost
of on-site work for the refurbishment of secondary
equipment. Prudent design of the substation plans
ahead for the mid-life refurbishment of secondary
schemes, when often the primary equipment is left
as-is, given that decades of years of serviceable life
may still remain. This permits the refurbishment
activity to take place with the absolute minimum of
primary system outage.
z
Improved communications capabilities
Data exchange between intelligent devices, intra and

inter-substation, is optimised through Ethernet
communications. Smart local and wide area control
units (WACU) can allow data exchange between
voltage levels in substations, and between

substations. Direct intersubstation communication
without the need to transit via a control centre
reduces the response times, for fast, real-time
applications.
3.
ARCHITECTURE OF DIGITAL
SUBSTATIONS
IEC 61850 is the international standard for Ethernetbased communication in substations. IEC 61850
allows for the full digitizing of the signals in a
substation and is necessary for the large amount of
data to be managed and communicated for the realtime management of a smarter, modern power grid.
IEC 61850 is designed for interoperability and
longevity, in order to be independent from any
supplier.
z
The process level
A digital substation is based on a communicating

architecture, whereby real-time operational
measurements are polled from the primary system.
This data is obtained using sensors, embedded within
the primary system, which is termed as the electrical
process. It is communicated to devices which must
act on those measurements by means of a process
bus. Most important is that smart devices and
systems within the substation can immediately
process this operational data. By subscribing as
clients to this data flow over an Ethernet process
bus, the information from the eyes and ears of the
power system is distributed and communicated
much more efficiently to the bay level than in
conventional hardwired schemes.
The process bus is also the link by which the primary
equipment information from out in the yard travels
back to the substation control house it links the
field back to the substation. In a fully digital
architecture, control commands (switchgear operator
commands, protection trips) also are routed to the
primary devices via the process bus, in the opposite
direction.
z
The protection and control level
Between the process bus and the station bus are

devices historically identified as the secondary
equipment. In the digital substation, these devices
are IEDs (intelligent electronic devices), interacting
with the field via the process bus, and with other
302

April 8-10
2015

Fig. 1: Architecture levels in a digital substation
peer devices in the bay, to other bays, and the digital

control system via the station bus.
z
The station control area
The digital substation station bus is much more than

a traditional SCADA bus, as it permits multiple clients
to exchange data, supports peer-to-peer device
communication, and links to gateways for intersubstation wide-area communication. GOOSE is
more often than not used as high-speed exchange of
binary status information/commands.
In addition to the need for distributed intelligence
between IEDs at the station level, there is the need
to distribute the information to local or remotelystationed control operators to visualize the
operational status of the substation.
4.
DIGITAL INSTRUMENT
TRANSFORMERS
The root of many of the limitations of conventional

instrument transformers is the reliance upon an iron
core. The core is a source of inaccuracy, due to the
need to magnetize it, but not to overflux it. In the
case of conventional CTs, achieving the low-level

accuracy and dynamic range to satisfy both
measurement and protection duties is a challenge.
Instead of an iron core, the translation from primary
to secondary measurement may use optical,
Rogowski or capacitive technology, with the
optimum choice for AIS (Air-Insulated Substation)
and GIS (Gas-Insulated Substation) driven by the
respective digital device size, which in turn permits
size optimisation of the switchgear. Some examples
of the principles are as follows, considering just the
CT function here for brevity:
z
303
Optical sensors (AIS) use the Faraday effect,

whereby a fibre optic loop sensor carrying a
polarised light beam encircles the power
conductor. This light will experience an angular
deflection due to the magnetic field, generated
by the primary current flow.
Rogowski sensors (GIS) dispense with the
conventional CT core and instead implement
windings as tracks on a multi-layer printed circuit
board. Four quadrants of the board are clamped
together to form a toroid around the primary

April 8-10
2015

conductor. The sensor output becomes a lowlevel voltage measurement, which can be
accurately correlated to the primary current.
merging units (such as shown in Figure 4) are

available, digitizing the CT and VT outputs at any
convenient kiosk out in the yard.
Digital controllers (SCU - switch control units) are
the fast, real-time interface to switchgear, mounted
close to the plant which they command. They replace
hardwiring of inputs/outputs by an Ethernet interface
to the yard. Where required, the AMU and SCU
function can be provided in a single integrated
merging unit, such as is shown below:
Fig. 2: Example optical CT showing freedom to mount in different

orientations, such as horizontally
Fig. 4: Example integrated merging unit for analog and digital

marshalling
6.
Fig. 3: Example flexible optical CT wrapped-around a bushing
5.
ANALOG SIGNAL CONVERSION,

MERGING AND SWITCHGEAR
CONTROL
Primary converters associated with each CT and VT

convert analog signals from the primary equipment
into digital signals. The primary converters interface
with merging units to perform all the digital data
processing necessary to produce a precise output data
stream of sampled values according to the IEC
61850-9-2 standard.
For retrofitting, or where the client has a preference
to retain traditional instrument transformers, analog
NUMERICAL PROTECTION RELAYS
In a fully-digital architecture, protection relays

receive currents and voltages as IEC 61850-9-2
sampled values, and issue trip or alarm signals using
IEC 61850-8-1 GOOSE. The authors companys IEDs
extend their supervision facilities to include
comprehensive addressing and plausibility checking
of the incoming sampled values from the process
bus. This addresses the fact that the traditional task
of current and voltage sampling is now external to
the device, and is connected via Ethernet. This
ensures maximum security, dependability and speed
of the protection scheme.
7.
DIGITAL CONTROL SYSTEM (DCS)
The DCS is the intelligence which binds together the

digital substation. It is central to the flow,
management and presentation of all components in
the digital substation. In the authors company,
specific focus has been paid to the way that
304

April 8-10
2015

operational and plant condition monitoring data has

been networked, for the first time, within what would
have traditionally been exclusively a protection and
control system. This avoids the need to overlay
multiple Ethernet networks, as the system is
deployed as a generic whole. In addition, attention
has been paid to how data can be presented as simple
dashboards, such that operational staff can clearly
see what is happening on the network, easing
subsequent decision-making on actions to take.
All digital substation architectures can be set up as
an IEC 62439 standards-compliant selfhealing ring
(HSR protocol) or dual-homing star (PRP protocol);
both of which are bumpless redundant. This means
that data is exchanged between devices via two
diverse paths, and should one of these paths fail,
data is instantly available hot from the other, with
zero delay. Fibre optic networks link all the systems
components, together and with the operator interface
(HMI), through a full range of Ethernet switches.
Wide area control units (WACU) offer the possibility
to exchange IEC 61850 GOOSE data between
Fig. 5: Example process bus protection relay
Fig. 6: Digital Control Systems: the marriage of state-of-the-art

software and intelligent electronic devices
voltage levels within a substation and also between

neighboring substations.
z
On-line condition monitoring and asset

management
Condition monitoring functions are mainly delivered

for power transformers, circuit breakers,
disconnectors and gas insulated switchgear. Physical
parameters are continuously monitored and real-time
measurements are combined and compared to
models in order to generate specific
recommendations regarding operation and
maintenance, as well as alarms when necessary. The
new architecture enables utility teams to have an
overview of the condition of all substations in real
time, and take appropriate and strategic asset
management decisions.
8.
ENERGINET, DENMARK EXAMPLE
In this project, the protected circuits are hybrid lines,

consisting of 400kV lines, and cabled portions laid
sub-sea and underground where required. There are
paralleled pairs of cables each 5km length, and the
operational demands are such that auto reclosing is
required for faults on the overhead lines, but not for
faults within the underground cable sections. The
differential protection is thus used for fast and
precise detection of faults within the cables. The
cables are a part of the two main 400kV connections
running from the south to the northern part of
Denmark.
The equipment supplied includes 72 optical CT units,
24 merging units and 24 line differential relays which
subscribe to the sampled values in a process bus
protection scheme. A lightweight dry type insulator
and window head design allow mounting of the
Optical CT (COSI) and CVT on a single pedestal
saving valuable yard space. For Energinet, Denmark,
a single structure and foundation per phase carries
the larger mass of the cable, plus the optical CT
support on a cantilever frame extending to 2 metres
horizontal distance. The reduced size and weight are
attractive benefits over conventional combined
units, allowing placement in compact substations
where space may be limited. Its broad dynamic range
makes this CT particularly suited for independent
merchant plants where extreme accuracy at both full
power output and station service draw is demanded.
In order to improve safety, the removal of wired
305

April 8-10
2015

cross-site current transformer circuits reduces the

risk of lethal injury due to inadvertent opening of
the circuit by personnel. The avoidance of oil in
instrument transformers reduces explosion risks too.
a patch box mounted inside the cubicle. Fibre

connections between P594 GPS time synchronising
units and merging unit/P546 current differential
relays are done directly via patch cables on the rear
of the panel in the 19 rack.
Fig. 7: Optical CT and Terminal Bushing on Single Support at

Energinet.dk Substation, Denmark
All protection relays (Alstom MiCOM P546 and

P594) and merging units are mounted in 19 racks.
Fibre connections from the cable management box
(CMB) in the yard to protection panel are spliced at
Fig. 9: Merging Unit and Process Bus Protection Relay Panel at
Energinet, Denmark
9.
Fig. 8: Topology Overview
TESTING PLATFORM
An Omicron test set was used to inject current

directly through the primary of the COSI-CTs. In order
to limit the required magnitude of current to be
injected, multiple turns were passed through the
COSI-CT. This enables the scheme to be tested as an
in-service simulation, proving the scheme from
primary current through to current differential
protection. Two main categories of tests were done
to prove the current differential protection, these
are the bias characteristic and the operating time.
The bias characteristic was plotted to show the
implementation of the sampled values has no impact
on the protection characteristic. To further prove the
operation of the protection scheme, the operating
times for internal faults were also submitted during
the FAT.
306

April 8-10
2015

Fig. 10: Diagrammatic Test Overview
10. CONCLUSION
Digital substation implementation allows the lifetime
total cost of ownership of the substation to be
reduced. The reduced size and weight of digital
instrument transformers, and the protection and
control panels provide attractive benefits, allowing
placement in compact substations where space may
be limited. The Energinet project shows the increasing
confidence in the application of digital substations
within Europe. This is essential at this network

voltage level, where cost savings and health and
safety are of paramount importance. As such, this
project is a valuable return on experience for other
utilities to follow in the same manner for both new
and refurbishment projects.
The authors wish to thank and acknowledge the
support and assistance of Energinet.dk in the frame
of this pioneering project.
307

April 8-10
2015

Long Term Experience with Natural Ester as Dielectric Fluid in

New and Retrofilled Transformers
Kevin J. Rapp
John Luksich
Alan Sbravati
Cargill
SUMMARY
Natural Ester insulating liquid has been used in
distribution and power transformers for many
years now. The experience in thousands of power
transformers with natural esters, in new and
retrofilled units, up to the largest 420 kV power
transformer, is traced back to the beginnings in
the early 2000s. Field operation data as well as
laboratory studies provided from power
transformers and reactors are discussed in this
paper to demonstrate that natural esters are a
reliable long term dielectric insulation and
coolant. In this paper, a case history and
laboratory testing data from samples collected
during routine maintenance intervals indicate that
natural ester fluid results can be used to reliably
diagnose operating problems and suggest
corrective actions.
Keywords: Power Transformers, Natural ester,
transformer insulation, retro-fill, transformer oil,
ester fluids, field experience with natural esters,
oxidation stability
1.
INTRODUCTION
Since being introduced for the first time in new

distribution transformers in 1996, natural esters have
been used for retro-filling since 1997 [1]. In the years
up to now, over 6,00,000 transformers have been
put in the field, delivering reliable service through
all the transformer sizes both in new and retrofilled
units. The number of small, medium and large power
transformers has grown to over 25,000 units. Some
of the experience of natural esters in power
transformers will be reviewed.
The combination of environmental, safety and
thermal capability to optimize the size of the
Sabine Bowers
Rajaram Shinde
transformer becomes an important driver. Figure 1

shows the progress of natural ester fluids from
discovery through mature utilization [2].
Fig. 1: Progress of natural ester fluid development and use
Whilst high fire point (more than 2x that of mineral

oils in transformers) is of high importance for some
users, the very positive environmental profile with
the highest biodegradability, the lowest CO2 footprint
and a sustainable resource as a raw material, natural
ester transformer oil is a sustainable future alternative
to mineral oil filled transformers of all sizes.
Laboratory and field studies have been conducted
to provide confidence of the increased capability &
acceptance of natural ester liquids [3-6].
A case study demonstrates that natural ester liquid
protects cellulose insulation allowing longer life and
increased thermal protection [7-10].
A summary of the natural ester-filled power
transformers will be provided in this paper, the
largest in operation being a 420 kV unit built by
Siemens in Germany for a German utility. In India,
308

April 8-10
2015

the largest unit built to date is a 75MVA 115kV

transformer by Transformers & Rectifiers for a
customer in Columbia, demonstrating that India is
geared up to adopt cutting edge proven technology
to enter the global market.
The paper presents data and analysis from a power
reactor as an interesting case study. This field study
provides highly useful information, giving confidence
that analysis of in-use natural ester liquids to diagnose
field performance is very similar to that with mineral
oil.
2.
IN SERVICE TRANSFORMERS WITH

NATURAL ESTER
Fig. 3: Power level of medium to large power transformers filled

with Natural ester liquid
2.1 Overview
The numbers of natural ester liquid-filled units that
are presented in this paper and the various
transformer types are based on cumulative data as
per our records. Not all transformers which have
been filled with Natural Ester can be included and
presented in this overview since it is impossible to
track every end-customer of transformers with this
fluid.
The majority of natural ester-filled power
transformers include over 25,000 small power
substation units above 38 kV and between 2 to 10
MVA in power from various transformer
manufacturers. Figures 2 and 3 provide a summary
of the voltage and power ranges between medium
to large power transformers containing natural ester
liquids.
In this overview of the medium to large power

transformers range, over 160 units are between 5
and 10 years old and nearly fifty units are more than
10 years old. The earliest new transformers were 69
kV 20 MVA units in the USA which have been
operating since 2003. The oldest known retrofill was
a 69 kV 50 MVA unit in 2001.
Together, the in service experience and field data
comes to a cumulative total of over 1000 years of
natural ester-filled new power transformer field
operation and almost 600 years for the retrofilled
units. The types of new and retrofilled medium to
large power transformers filled with natural ester
liquids are presented in Table 1.
Fig. 2: Voltage level of medium to large power transformers filled

with Natural ester liquid
309
Table 1: Types of natural ester filled power

transformers

April 8-10
2015

2.2. First 420 kV Power Transformer

In early 2014, TransnetBW, a German transmission
network operator, commissioned a Siemens 420 kV
extra-high voltage transformer filled and shipped
with Envirotemp FR3 Fluid, natural ester
insulating liquid. The transformer shown in Figure 4
was installed at a substation in Bruchsal to link the
380 kV extra-high voltage level with the 110 kV grid.
The rated power of the 420 kV transformer is 300
MVA with continuous loading capacity up to 400
MVA, which takes advantage of the higher thermal
class of thermally upgraded paper immersed in
natural ester liquid. The design and cooling
characteristics of the transformer in accordance with
standard IEC 60076-2 are expected to provide a
longer life span than a similar mineral oil unit [11-12].
The transformer had a special partial discharge
requirement of less than 10 pico-coulombs at 1.1
p.u. of rated voltage that was successfully achieved
at factory acceptance testing.
The insulation system was designed in accordance
with IEC 60076-14 for liquid immersed power
transformers using high temperature insulation
materials [2,13]. The combination of thermally upgraded
Kraft paper and natural ester liquid provides a
significant advantage from the temperature limit of
130C on the winding hot spot. Table 2 summarizes
the design characteristics of the transformer.
Fig. 4: Siemens 420 kV extra-high voltage power transformer

filled with Natural ester liquid
2.3 Largest Transformer with Natural Ester

in India
Following hot on the heels of the successful
introduction of natural ester in the highest voltage
rating in Germany, it took only a few months for
Indian manufacturer Transformers & Rectifiers
(T&R) to adopt & develop this new age technology.
The first 75MVA 115kV Power transformer
filled with Natural Ester Envirotemp FR3, was built
in India by T&R in Ahmedabad for a Columbian
customer.
Table 2: Design Characteristics of Siemens 420 kV

transformer [2, 11]
Fig. 5: 115 kV 75 MVA Power transformer filled with Natural ester

liquid by Transformers & Rectifiers India
The 420 kV transformer is equipped with a MR OLTC

VACUTAP VR.
Further to this success journey in India, some more

companies accelerated this process of futuristic
310

April 8-10
2015

technology adoption like eg Schneider, ABB,

Raychem RPG etc.. transformers of 20 MVA 33/
11kV ratings are currently being built by Schneider
for the Indian Utility Tata Power in Mumbai,
optimized around the higher temperature profile
according to IEC 60076-14 to allow a smaller
transformer with a higher rating. The 2 transformers
are designed to operate at a permanent capacity of
28MVA at a smaller footprint to fit into the space on
a prestigious commercial hub in Mumbai.
more than 7 years. The map in Figure 6b indicates

the climate conditions of the reactor installation
region: hot and wet throughout the year (no dry
season).
2.4 Reactor Case Study

Collaboration between customer of the transformer,
transformer OEM and Cargill has proven to be of
important benefit to all when looking at the following
case study. A three phase shunt reactor rated at 138
kV / 11.4 MVAr was designed, filled and tested in
factory with Envirotemp FR3 natural ester liquid.
Eletrobras/Eletronorte Grid Company energized the
reactor in 2006. The customer was advised to have
regular checks as for all other equipment. The
customer found some changes in the Ester fluid
behavior during the 4th year of operation & continued
with monitoring the behavior. [14]. Testing results
indicated a very atypical variation of the fluid
parameters, in comparison with other natural ester
liquid-filled equipment owned by the company.
Figure 6a is a graph of the neutralization number of
several FR3 natural ester liquid filled transformers
and reactors at Eletronorte. The graph clearly shows
the Deviation.
Fig. 6a: Acid value of natural ester-filled transformers and

reactors in service for about 7 years
This was the right opportunity for the Cargill team

in collaboration with customer to further investigate
the behaviour. Active collaboration to understand
long-term behavior of the natural ester in normal
and abnormal conditions is critical to acceptance of
this technology.
In September 2014 an on-site inspection of the
reactor was conducted and the cause of the behavior
change was identified: Improper rubber bag
installation.
The attachment point of the bag in the upper flange
of the oil conservator was not properly assembled,
allowing free entry of atmospheric air into the unit.
Simply, the natural ester oil immersed reactor was
in a completely free breathing condition. This means
the Unit was operating as free breathing, unlimited
access to air and moisture in extreme conditions for
Fig. 6b: Brazilian climate map showing the location where the
reactor is installed
The reactor remained energized at rated voltage

continuously (full load condition which is typical
loading of reactors) during the seven years it has
been in service. Figure 7 confirms that the outside
of the rubber bag has been in open contact with
atmospheric air, clearly visible as a horizontal line
of the oil level. The air was occupying the upper
volume outside the bag.
311

April 8-10
2015

Fig. 7: Outside of the rubber bag, where the separation line

indicating the "oil level" can be clearly seen
Application of natural ester liquids in free breathing

equipment is not advisable and what happened with
this reactor was unintentional. But it case has given
huge inputs on an actual in-field performance of
natural ester liquids in free breathing conditions. It
is important to consider that the reactor has been
operating as free breathing condition in a region of
very high moisture (it rains everyday), very hot
(ambient temperature between 30 and 40C), under
severe thermal condition (24/7 full loading).
According to the records, top oil temperature was
typically around 90C or higher.
The equilibrium condition of the natural ester liquid
exposed to a hot wet climate should result in a
relatively high moisture condition. However, the
maximum moisture content measured in the
liquid during the 7 years was 183 ppm, as shown in
Figure 8.
Fig. 8: Total water content in the natural ester during

7 years in service
Fig. 9: Kinematic viscosity at 40C and 100C of the natural ester

liquid during 7 years of field service in the 138 kV reactor
Assuming a conservative average oil temperature of

70C, the water saturation level of FR3 fluid would
be about 2263 mg/kg, which results in 8.1 %
maximum relative moisture content. Despite the
existence of a silica gel absorbent that helped reduce
the moisture ingress, this level is surprisingly low,
being a very clear indication that the moisture
absorbed from the atmosphere was consumed by a
hydrolysis reaction with the natural ester liquid. The
acid by-products of hydrolysis increased the
neutralization value, but kept the fluid dry. As a
result, the paper was also kept dry, even under very
adverse environmental conditions.
During the on-site inspection a procedure for
measuring the insulation moisture was performed.
The tank was kept under vacuum for a few hours
and then under dry air (dew point lower than -50C)
during 48 hours. The measurement of the dew point
of this air in equilibrium with the inner tank
condition is used to indicate the moisture content
of the insulation. The measured dew point of -25C
for an ambient temperature of 31C indicates about
0.90 to 1.05 % of moisture in the solid insulation
inside the reactor. For this 138 kV unit, this moisture
content is close to the requirements for new
equipment.
This is the most important aspect of this case study,
since it confirms that the natural ester liquid, despite
having a very high availability of moisture from the
ambient, was still providing moisture extraction from
the solid insulation. Instead of equilibrium between
the insulating paper and liquid, in this case the
312

April 8-10
2015

equilibrium included the outside ambient conditions.

The liquid was strongly affected as the acid value
shows, but the ambient moisture conditions did not
really affect the solid insulation, which was in fact
drier than would be expected in a heavily loaded
mineral oil unit.
The dry-out effect [15] of the fluid to extract moisture
from the paper and reduce the thermal degradation
rate, in this case, was not cancelled by exposure to
the ambient conditions. In fact, the long-chain acids
produced by the hydrolysis reaction between water
and the natural ester liquid is a key concept that is
different compared to mineral oil. Short-chain acids
from mineral oil catalyze the thermal breakdown of
cellulose versus the long-chain acids from natural
ester liquids provide increased stability to the
cellulose via a transesterification reaction hypothesis
described in [7] and [8].
Figure 9 shows a slight increase in viscosity at 40C,
but not seen at 100C. This increase is not enough
to indicate a critical oxidation of the natural ester.
The increase of the neutralization value, as a result
of the moisture ingress, was much faster than the
possible oxidation caused by the continuous contact
with ambient air.
From Figure 5, taking into account that the suggested
limit for neutralization no. for continuous operation
of equipment is 0.3 mg KOH/g, an investigation
should have been started in the middle of 2010.
However, since the equipment was being
continuously monitored and the dielectric strength
shown in Figure 10 was not indicating any problem,
no action was taken.
Fig. 11: Dissipation factor of the natural ester liquid from the
reactor, with the red line indicating the suggested limit of
continuous operation as per IEEE C57.147 [16]
In late 2011, the dissipation factor investigation limit

was surpassed, as can be seen in Figure 11. Likewise,
the decision was taken to continue operation and
monitor. The dielectric breakdown voltage decreased
slightly in late 2011, but this was probably test
variability, as it increased in subsequent sample
measurements.
The main conclusions drawn from the 138 kV reactor
case study are:
z
The increase of the neutralization no. due to

natural ester hydrolysis with the ambient
moisture is much faster than the effect of
oxidation on the fluid viscosity.
Despite not measuring the oxidation inhibitor

content, the stability of the viscosity indicates
that seven years while under the severe
conditions was not enough to consume all
oxidation inhibitor.
Fig. 10: Breakdown voltage of the natural ester from the reactor,
with the red line indicating the suggested minimum value for the
fluid after contact, prior to energization (new)
313
The fluid behaved as a sacrificial component

from the free-breathing and moisture caused by
the incorrect installation, however, the
degradation did not affect the dielectric
performance of the fluid.
It is not possible to estimate exactly, but the
solid insulation moisture level expected for a
reactor under similar conditions filled with
mineral oil would be higher than what has been
identified in this case.
The soy-based natural ester insulating liquid is
robust for real world applications, even when
problems develop.

April 8-10
2015
3.

CONCLUSION
symposium 2014, Filderstadt, March 12, 2014.
Natural esters transformer dielectric oils are moving

fast towards the acceleration phase. After proving
excellent performance in distribution transformers
(up to 36 kV) for many years, the voltage and power
have slowly increased, reaching the 420 kV and 400
MVA level last year, indicating fast growing
confidence. India has lately seen adoption of this
fluid both for internal and external customers,
demonstrating that the Indian market is aggressively
adopting this new age technology faster than others
to benefit to the nation in green movement, fire
safety, increased efficiency & value for money. .
Currently, there are a number of 500 kV projects
and higher under study.
The natural ester liquid-cellulose interactions
identified in the Free-breathing reactor case study
in Section 2 and presented both in IEC [13] and IEEE
[17]
standards have been the basis for several utilities
to increase the temperature rise limits of their
transformers. The reduction of the paper degradation
rate for the same temperature has a very positive
financial effect on asset depreciation and net present
value.
Additionally, this new and improved technology has
been instrumental in equipment cost optimization,
size and material reductions and improved thermal
capability of electrical power equipment, which has
added momentum to the adoption of natural ester
filled transformers by utilities.
There are significant numbers and ages of natural
ester filled transfrmers and reactors in service around
the world without any critical issues caused by the
liquid. This is the strongest evidence that natural
ester liquids are a very robust and reliable solution
as an improved liquid insulation for transformers and
other electrical devices.
BIBLIOGRAPHY
1.
J. Luksich, K. Rapp, Suitability of Natural Ester

Fluids for Retofilling Transformers, Weidmann
2012 Technical Conference, Las Vegas, NV
October 18, 2012.
2.
M. Schfer, I. Atanasova-Hhlein, R. Fritsche,

C. Schmeid, Use of Natural Esters in Power
Transformers Operation Experiences and State
of the Art, Stuttgarter Hochspannungs
3.
C.P. McShane, G.A. Gauger, J. Luksich, Fire

Resistant Natural Ester Dielectric Fluid and
Novel Insulation System for its Use, IEEE/PES
Transmission & Distribution Conf., April 12-16,
1999, New Orleans, USA
4.
Jerry Murphy, Rob Weber, New Coolant

Improves Transformer Performance,
Transmission and Distribution World, Feature
Article, September 2004
5.
David Bingenheimer, Luiz Franchin, Eugene Del

Fiacco, Jos Mak, Vagner Vasconcellos, Kevin
Rapp, Sustainable Electrical Energy Using
Natural Ester Technology, CIRED 21st
International Conf. on Electricity Distribution,
paper 1148, June 6-9, 2011, Frankfurt, Germany
6.
Kevin Rapp, Measuring Acids to Monitor the

Health of Vegetable Oil-Filled Transformers,
American Chemical Society Pittsburgh Conf.,
March 21, 2013, Philadelphia, USA
7.
C.P. McShane, K.J. Rapp, J.L. Corkran, G.A.

Gauger, J. Luksich, Aging of Paper Insulation
in Natural Ester Dielectric Fluid, IEEE/PES
Transmission & Distribution Conf., October 28November 2, 2001, Atlanta, USA
8.
K.J. Rapp, C.P. McShane, J. Luksich, Interaction

Mechanisms of Natural Ester Dielectric Fluids
and Kraft Paper, IEEE/DEIS 15th International
Conf. on Dielectric Liquids, June 26- July 1,
2005, Coimbra, Portugal
9.
D. Martin, Z.D. Wang, I. Cotton, The Use of

Natural and Synthetic Ester-Based Transformer
Oils in Power Transformers, Proceedings of the
XIV International Symp. On HV Eng., Tsinghua
Univ., Beijing, China, August 25-29, 2005
10. D. Martin, Z>D> Wang, A.W. Darwin, I. James,

A Comparative Study of the Chemical Stability
of Esters for Use in Large Power Transformers,
IEEE/DEIS Conference on Electrical Insulation
& Dielectric Phenomena, October 15-18, 2006,
Kansas City, USA
11. Ronny Fritsche, Uwe Rimmele, Frank
Trautmann, Michael Schfer, Prototype 420 kV
Power Transformer Using Natural Ester
Dielectric Fluid, TjH2b-19th Annual High
Voltage Electrical Equipment & Management
314

April 8-10
2015

Conf., TechCon North America, Feb. 3-6, 2014,

Phoenix, USA
12. IEC 60076-2 Ed. 3, Power Transformers Part
2: Temperature rise for liquid-immersed
transformers, International Electrotechnical
Commission, Feb. 2011, Geneva, Switzerland
13. IEC 60076-14 Ed. 1, Power Transformers Part
14: Liquid-immersed power transformers using
high-temperature insulation materials ,
International Electrotechnical Commission, Sep.
2013, Geneva, Switzerland
14. Iran P. Arantes, Marcelo Martins, Joo Baldauf,
Andress Baccin, Ito Capinos, Shunt Reactor
with Vegetable Oil Insulating Fluid Design
Consideration, Tests and Field Experience,
CIGRE-Brasil, VI Workspot International
Workshop on Power Transformers, April 25-28,
2010, Foz do Iguau, Brazil
15. Steve Moore, Kevin Rapp, Ramona Baldyga,

Transformer Insulation Dry Out as a Result of
Retrofilling with Natural Ester Fluid, IEEE
Power & Energy Society, Transmission and
Distribution Conference & Exposition, May 710, 2012, Orlando, USA
16. IEEE C57.147-2008, Guide for Acceptance and
Maintenance of Natural Ester Fluids in
Transformers, Institute of Electrical and
Electronic Engineers Power & Energy Society,
July 2008, New York, USA
17. IEEE C57.154-2012, IEEE Standard for the
Design, Testing, and Application of LiquidImmersed Distribution, Power, and Regulating
Transformers Using High-Temperature
Insulation Systems and Operating at Elevated
Temperatures, Institute of Electrical and
Electronic Engineers- Power and Energy Society,
Oct. 2012, New York, USA.
315

April 8-10
2015

Digital Substation Understanding the Technology, Initiatives and

Future Planning in GETCO
R.P. Satani
Gujarat Energy Transmission Corporation Limited (GETCO), India
SUMMARY
In a conventional substation the primary
equipments i.e. instrument transformers, circuit breaker, disconnectors etc. are connected to
control and relay panels / bay controllers through
hard wiring - copper control cables. This requires
much labour and it is time consuming job. It is
also very difficult to identify the mistakes / faults
in wiring. These all together ultimately contribute
to long project delivery time. The primary
equipments themselves are also not having any
intelligence. Some inherent issues associated with
conventional primary equipments can not be
eliminated. For example, the conventional CT is
still with the probabilities of secondary open
circuit, saturation, inability to produce exact
replica of primary current particularly high
frequency wave forms/harmonics. Complete
monitoring of circuit breaker and disconnectors
further adds cabling and separate devices. As far
as conventional secondary equipments,
particularly protection system, is concerned, it is
required to share the information among the relays
which again need inter relay hard wiring if relays
are with different protocols/standards.
The digital substation is the solution to address
the above issues. Also, to transform real time data
and control to master control centre and to reap
maximum benefits of information technology and
digital technology in control, protection and
automation, the digital substation is todays need
to keep pace with the changing power system
scenario.
Since it eliminates hard wiring, the operational
and maintenance cost will also be lower and thus

it gives long term benefits in terms of Maintenance
and Life Cycle Cost.
This paper explains about what the digital
substation is and what initiatives GETCO has
taken to move towards digital substation in India.
Keywords: Digital Substation, Optical CT, Merging
Unit, Intelligent Control Unit, IEC 61850 - 9-2,
Process Bus
1.
DIGITAL SUBSTATION AND ITS

DIFFERENT TYPES
In digital substation, the primary equipments are

connected to control and relay panels / bay
controllers through fibre optic cable. Primary
equipments like circuit breaker, disconnectors are
made intelligent by providing Intelligent Control Unit
(ICU) while instrument transformers by Merging Unit
if they are conventional. Optical CT and Electronic
VT will also be used instead of conventional one.
Digital substation can be of three types as shown
below.
IEC 61850
adopted at
station
level
Scheme 1
Scheme 2
Scheme 3
2.
z
deengg.getco@gebmail.com
316
IEC 61850
adopted at
process
bus
Optical
CTs and
Electronic
VTs
WHY DIGITAL SUBSTATION? SALIENT FEATURES / BENEFITS

No hard wiring - Tremendous reduction in
cables and hence Cable Cost, Reduced Cable
trenches.

April 8-10
2015

Saving in erection, testing & Commissioning

time - Fast project delivery
Optical CT & Electronic VT - light in weight,

Improved measuring accuracy
Reduced Auxiliary contacts of Switchgears Reduced LCC size of Switchgears
Reduced Auxiliary Relays - Reduced size of C&R

Panel
Reduced command response time
Reliable Monitoring of Equipment - Higher

Transmission System Availability
a separate unit - installed in control panel put

in control room/yard kiosk. The Fibre Optic
cable is used to connect CT and merging unit to
carry the light (optical signal). Depending on
the current flowing though the conductor this
light beam rotate by some angle in proportional
to the current. This deflected light beam returns
to electronic circuit/merging unit which detects
the deflection and gives current signal output
in digital form compatible to the standard IEC
61850 - 9-2 for protection circuit, while in
analogue form with the help of amplifier unit
for metering circuit.
Least maintenance and manpower requirements
Salient Features / Benefits of Optical CT
Reliable and Accurate
No open secondaries - Safe
Avoids mal operation
Quick and correct data capturing, computation

and monitoring
No hard wiring - copper cables between

CT and Control panel
Dry Type, No Oil, No SF6 - No violent

failure, No flaming, No explosion risk - Safe
and Environment Friendly
No Saturation - Eliminates possibility of

mal operation of relays
Accurate Measurement as it dose not

affected by stray field
Wide Dynamic range - Up to 0.2 % to 150%

of rated current for Metering range while
up to 171 KA for Protection range
Wide Bandwidth - Reproduction up to 6

kHz
Can measure more accurately the highorder harmonics, transient current, DC

current
Operations closer to system element capacity
3.
DIGITAL SUBSTATION EQUIPMENTS
3.1 Instrument Transformers

(i)
Optical Current Transformer

It is a non conventional type Current
Transformer, which uses optical current sensor,
works on magneto-optic effect known as
Faraday Effect i.e. when a polarized light beam
passes through the electromagnetic field, it will
deflect by angle proportional to the intensity
of the electromagnetic field as depicted below.
z
z
No ferromagnetic resonance
Reduced size & Light Weight - Very
compact - Requires less foot print and
reduces erection time
Composite Insulator - Seismic Stable
Requires no maintenance
(ii) Electronic VT
In this CT, the ring type current sensor
(surrounds the current carrying conductor)
consists of few turns of optical fibre which
carries polarized light generated by the
Electronics circuit (also called Merging Unit) -
317
The electronic VT is a capacitor divider but

instead of intermediate electromagnetic
transformer as in the case of CVT, an electronics
converter is used. The low power signal from
this electronics converter is given to merging
unit which converts it into process bus IEC

April 8-10
2015

61850 9-2 compatible digital signal which is sent

to protection and control IEDs.
3.2 Circuit Breaker with Intelligent Control

Unit
An intelligent electronics device known as Intelligent
Control Unit (ICU) is installed in circuit breaker local
control cubicle (LCC) which interfaces with circuit
breaker through hard wiring - i.e. copper wires to
collect breaker data and communicates with control
& protection IEDs on IEC 61850-8-1 through FO
cable to control and monitor the breaker.
Tells about the mechanical service life of

disconnector - counts the numbers of
disconnector operations.
Control the movement of the disconnector

during opening and closing operations
3.4 Relays IEC 61850 9-2 Process Bus

compatible
Monitoring of circuit breaker operation

characteristics open / close status, Tripping/
closing coil current, time etc.
Unlike conventional numerical relays which accepts

analogue current and voltage signals, the relays used
in digital sub station receives sampled values of
currents and voltages in digitized form compliant to
IEC 61850 9-2 standard from merging unit which
merges current and voltage signals got either from
non conventional Instrument transformers like
optical CTs and Electronic VTs or conventional
electromagnetic CTs and VTS.
Supervision of spring charging motor circuit and

current
3.5 Substation Automation System
Tells about the mechanical service life of circuit

breaker - counts the numbers of circuit breaker
operations. Also measures contact travel and
speed.
Tells about the electrical service life of circuit

breaker - estimates the remaining electrical
service life of circuit breaker through the
monitoring of primary current.
Following functions are integrated in this device.
Acquisition and supervision of SF6 pressure,

temperature and moisture with the help of
sensors
3.3 Disconnectors with Intelligent Control

Unit
As in the case of circuit breaker, an Intelligent
Control Unit (ICU) is installed in disconnector
(Isolator) operating box which interfaces with other
devices/components in disconnector (Isolator)
through hard wiring - i.e. copper wires to collect
disconnector (Isolator) data and communicates with
control & protection IEDs on IEC 61850-8-1 through
FO cable to control and monitor the disconnector
(Isolator).
It is a system to control and monitor the substation

from a single operating console. It consists of Bay
Control Units (BCU), protection IEDs (Intelligent
Electronic Devices), fibre optic communication
network with ethernet fibre switches. Substation
automation software is heart of the system.
3.6 Time Synchronizing Equipment

All the events / disturbance records shall be time
stamped so as one can know what happened at which
moment and correct analysis can be made based on
these time tagged data. Many events / records are
required to analyze an occurrence and these data
are to be taken from more than one device and hence
all these devices shall be time synchronized which
requires a common reference time signal. GPS based
time synchronizing equipment is used for it. All the
devices synchronized either by using IRIG B or SNTP
protocol.
As far as the process bus devices merging units
are concerned, they will get time synchronization
signal on IEEE 1588 protocol.
4.
Following functions are integrated in this device.

z
Monitoring of disconnector operation

characteristics open / close status, motor
current for Opening/closing operation, time etc.
DEPLOYMENT BY UTILITIES - GETCO

INITIATIVES
(i) Optical CT
GETCO has installed Alstom make 220 KV optical
CTs in one bay - Achhalia Line No.3 at existing 220
318

April 8-10
2015

kV Jambuva substation and commissioned it on

dt.21.06.12. It is the first installation of its kind in
AC system in India. It is put in series with existing
conventional CTs. The Optical CT Panel, installed in
control room comprises Merging Unit, Current
Amplifier Unit and Separate distance protection relay
(MiCOM P444 - compatible to standard IEC 61850 9-2) and energy meter installed is connected to this
Optical CT to evaluate its performance
independently. As the substation is conventional
substation, a SCADA system is also installed in this
panel for this particular bay.
One such identical set of optical CTs is also installed
in 220 KV Godhra line bay on dt.22.06.12 at green
field 220 KV Lunawada substation.
Schematics / Architecture
Site Photographs
220 KV
Optical
CT
Performance
The optical CTs are working satisfactorily since

commissioning. The Optical CT and relay (MiCOM
P444) connected to it performed well and operated
correctly during faults on the line. Its disturbance
record is also matching perfectly with that of existing
relay (MiCOM P442) connected to conventional CT.
The operational performance of Jambuva Optical CT
project is tabulated as below:-
319
Optical Current Transformer at GETCOs 220 KV Jambuva

substation

April 8-10
2015

220 KV
Optical CT
is mounted in separate cubicle - suitable for

outdoor installation conforming degree of
protection IP-66
Cubicle consisting merging unit is erected in

yard near CT,CVT JB/Bay Marshalling Kiosk (BMK)
Intelligent control units

220 KV
Conventional CT
interface with CB and isolators through hard

wiring ( List of I/Os is given in Annexure -1)
communicates to control & protection IEDs on

IEC 61850 -8-1 through FO cable as mentioned
above at D2 & D3.
is mounted in separate cubicle - suitable for

outdoor installation conforming degree of
protection IP-66
Cubicle consisting merging unit is erected in

yard near CT,CVT JB/Bay Marshalling Kiosk
(BMK).
The architecture of MU and ICU for 400 KV Kosamba

line at 400 KV Asoj substation is as follow:-
Optical Current Transformer at GETCOs 220 KV Lunawada

substation
(ii) MU for conventional CT & VT and ICU

for CB & Isolators
As the cost of Optical CT is quite high with very few
suppliers, GETCO demanded a solution from
technology providers to convert analogue data of
conventional CT & VT into digital form. This will
further reduce nos. of cores in the conventional
instrument transformer as we would be using only
one core to split into multiple digital outputs.
However, metering core could be exclusive.
Accordingly, Merging Unit for conventional CT &
VT and Intelligent Control Unit for circuit breaker
and isolators in one 400 KV feeder bay at existing
400 KV Asoj substation are installed by M/s Alstom
T & D India Ltd. and same has been commissioned
in July 2014.
The merging unit
take analog inputs from CT & CVT
is compatible to existing conventional CT & CVT
converts these analog signals into digital outputs
Sampled values conforming to IEC 61850
9-2
communicates simultaneously to IEDs through
fibre optic cable over process bus
This is a beginning for adopting full fledged digital

substation technology with process bus in future.
5.
FUTURE PLANNING FOR FULLY

DIGITAL SUBSTATION
On satisfactory performance, we have decided to go

for one fully digital substation which comprises of
conventional as well as optical CTs, electronic VT,
320

April 8-10
2015

Intelligent control units for circuit breaker and

isolators, process bus compatible relays in all bays.
Seamless integration of Digital Substation with

System Operation/Load Despatch for its fullest
use
Integration with on-line diagnostic monitoring

equipments for substation operation
7.
CONCLUSION
The typical architecture of fully digital substation:
While the good amount of work have been done in

substation automation with IEC 61850 compliances
and time stamping of real time data, there is still
further scope to improve the automation system by
making full use of latest state of art Information
Technology, Communication Technology, Digital
Technology and Softwares to achieve accuracy of
data and reliable system operation. Such investments
have to save the cost in long run in terms of system
availability / uninterrupted power supply, fast
restoration and minimum maintenance and
manpower cost.
Digital Substation
6.
EXPECTATIONS
Outdoor application Fully compliant to Indian

climatic conditions?
Availability of Energy meters particularly tariff

meters compatible to Optical Instrument
Transformers
IEC 61850 - 9.2 compliant Bus Bar Protection Not available even with many of the leading
Relay manufacturer
Full use of Communication technology to spread

its functions for operation like SMS, email
We have to further advance substation automation

system by integrating online diagnostic data with
real time system operation and guide the utility for
proactive actions / measures. An exclusive research
work is required in this area.
BIBLIOGRAPHY
1.
Optical CT / Digital substation technical leaflets

of M/S Alstom and M/s NR Electric
2.
GETCO Jambuva, Lunawada and Asoj projects.
321

April 8-10
2015

List of ICU Inputs and Outputs

Analog Merging Unit (MU) for conventional CT &
VT and Intelligent Control Unit (ICU) for CB &
Isolators at 400 kV GETCO Asoj Substation
Sl. No.
Analog Voltage Inputs
Bus - 1 R-Ph voltage
A.
Bus - 2 R-Ph voltage
Line R-Ph voltage
List of I/Os for Circuit Breaker Intelligent

Control Unit
Annexure-1
Sl. No.
Digital Inputs
CB R-Ph Open
Sl. No.
Analog Inputs (4-20mA)*
CB R-Ph Closed
CB R-Ph Motor current
CB Y-Ph Open
CB Y-Ph Motor current
CB Y-Ph Closed
CB B-Ph Open
CB B-Ph Motor current
CB B-Ph Closed
CB R-Ph Closing Coil current
CB - Local Mode
CB Y-Ph Closing Coil current
CB - Remote Mode
CB B-Ph Closing Coil current
CB - AC supply Healthy
CB R-Ph Tripping Coil - 1 current
10
CB - DC supply - 1 Healthy
CB Y-Ph Tripping Coil - 1 current
11
CB - DC supply - 2 Healthy
CB B-Ph Tripping Coil - 1 current
12
CB - Spring Charged
10
CB R-Ph Tripping Coil - 2 current
13
CB - Lockout - 1 Optd
11
CB Y-Ph Tripping Coil - 2 current
14
CB - Lockout - 2 Optd
12
CB B-Ph Tripping Coil - 2 current
15
CB - PDR Optd
13
CB R-Ph SF6 gas Pressure
16
CB - Motor MCB trip
14
CB Y-Ph SF6 gas Pressure
17
CB - Ready
15
CB B-Ph SF6 gas Pressure
18
CB - SF6 Gas Press Low
Sl. No.
Digital Outputs
CB Open command - 1
CB Open command - 2
CB Close command
Sl. No.
Digital Inputs
4, 5
CB - TC - 1 R-Ph Trip (Protn + Manual)
Isolator R-Ph Open
6, 7
CB - TC - 1 Y-Ph Trip
Isolator R-Ph Closed
8, 9
CB - TC - 1 B-Ph Trip
Isolator Y-Ph Open
10, 11
CB - TC - 2 R-Ph Trip
Isolator Y-Ph Closed
12, 13
CB - TC - 2 Y-Ph Trip
Isolator B-Ph Open
14, 15
CB - TC - 2 B-Ph Trip
Isolator B-Ph Closed
* - At present, it is not connected but provision is kept

in ICU
322
List of I/Os for Isolator Intelligent Control

Unit

April 8-10
2015

Isolator - Local Mode
Sl. No.
Digital Outputs
Isolator - Remote Mode
Isolator Open command
Isolator - AC supply Healthy
Isolator Close command
10
Isolator - DC supply Healthy
Isolator I/L
11
Isolator - R-Ph O/L relay Optd
12
Isolator - Y-Ph O/L relay Optd
Sl. No.
Analog Inputs (4-20mA)*
13
Isolator - B-Ph O/L relay Optd
Isolator R-Ph Motor current
14
Isolator - PDR Optd
Isolator Y-Ph Motor current
15
Isolator - Manual Mode
Isolator B-Ph Motor current
16
Isolator - Motorized Mode
* - At present, it is not connected but provision is kept

in ICU
323

April 8-10
2015

Generating Useful Transformer Health Indices

A.J. McGrail
Doble Engineering, USA
S. Gaikwad
Doble Engineering, India
SUMMARY
This paper looks at transformer health indices,
and the generation of an index so as to quickly
provide value to an organization. A transformer
health index is a number which represents the
health of a transformer it can be used to rank
transformers for replacement, to identify those
most likely to fail, to identify those which can be
left alone.
There are many ways to generate an index, some
of which are more useful than others. The key is
to understand that we are making decisions about
a large asset, but we are generating condition data
which may be indicative of a particular condition,
and that to link the condition data to the asset we
have to address failure modes. If the generated
condition data does not indicate a failure mode,
it may be of limited value. If it does indicate a
failure mode, how strongly is that failure mode
indicated, and how confident are we in the
diagnosis? Relevance of a data source to a failure
mode is useful to determine whether a particular
mode is active, based on several data sources.
Getting use from a health index should not require
every possible piece of data building a system
which can start with limited data sets, such as off
line DGA alone, and then build up by adding in,
for example, test and maintenance data, or asset
specific history, is a means to provide value early,
and to get user feedback on asset health systems.
This paper will describe the generation of health
indices from a simple Delphic approach, through
basic DGA analyses on to complex statistical
approaches based on relevant standards; in each
case practical examples are given and discussed.
sgaikwad@doble.com
Keywords: Transformer asset health index Delphi

DGA statistics.
1.
WHAT IS A HEALTH INDEX?
A condition or health index of an asset is an approach

to identify those assets most likely to need some
form of intervention based on condition: whether
refurbishment, replacement or maintenance or some
other form of activity.
We can ask a simple but profound question:
What does an asset health index or asset health score
actually mean ?
Consider a scale of 1-10, where 1 is as new and 10
is close to failure. Does a score of 6 out of 10 mean
that we have used 40% of the asset life? That we
can load it to 60% of nameplate? That it is likely to
fail with a probability of 60% in the next year?
It may be intended is that a 6/10 score for an asset
means that it is less likely to perform its function
successfully than a 5/10 asset. Generating scores to
several decimal places doesnt reflect the true
uncertainty in an analysis, but we often see such
results.
Condition ranking may include data from several
sources, each of which may be assessed and ranked:
Maintenance and repair activities
Off line testing results
Inspection data
Online monitoring data
Operational data load, tap position etc
Expert input
Industry information on designs
How we deal with incomplete data will have a
significant impact on the viability of the scoring
system. Combining scores from several sources is a
possibly complex activity.
324

April 8-10
2015

The simple question, posed above, can be extended

to give it more power:
4.
What does an asset health index mean in terms of:

-
what I need to do?
how urgently do I need to do it?
These points are key to getting value from an asset

health index each score or range of scores needs to
have a timescale, to indicate urgency, and a range of
actions depending on the asset in question and the
source data.
2.
METHODS OF GENERATING AN INDEX
There are many methods available for generating an

asset health index we will describe some systems
here which can be used stand-alone or combined
into an overall score.
2.1 Delphic Approach
timescale
transformer is on active list for replacement
within 2 years
This is a condition based approach to asset

replacement; it is a strategic approach which will be
modulated by tactical contingencies as transformers
which are not in group 1 fail and need replacement.
The approach thus needs to be augmented with an
appropriate number of spare units a separate
analysis.
We also have to define the word last in the context
of each class this can be done by an individual
organization, but in the context of the discussion
here, last means perform its required function
without additional maintenance, monitoring or
intervention over and above that which is considered
normal or standard.
Each transformer is then given an evaluation score.

The score reflects both a timescale and an action
which are appropriate for the analysis in question.
The action is most easily identified as replacement.
This SME Delphic Approach has been used in practice

at a large T&D organization. The following steps were
used:
Bring SMEs together.
Agree terms of reference: definitions, timescales
etc.
Rank each transformer identify any anecdotal
evidence.
Review outcome as sanity check; iterate if
necessary.
Initiate gathering of corroborative data.
Share analysis with field and operational staff
for confirmation, discussion and revision.
Monitor population.
Transformers which are in need of replacement

require corroborative data and analysis from
maintenance, off line test, online monitoring or
operational data. This confirms that the Delphic
approach as a getting started approach.
The approach worked - a population of >2,000

distribution and transmission transformers was given
an initial classification, subsequently modulated by
field staff feedback, and corroborative evidence
subsequently gathered for future reference.
As an example we may have scores 1 through 4:

1. transformer is in a condition which means it is
expected to last for the foreseeable future, and
at least 15 years
2. transformer is expected to last up to 15 years
but may need replacing on a 5-15 year
timescale
3. transformer is expected to last up to 5 years
and may need to be replaced in a 2-5 year
2.2 DGA Approach
In a Delphic approach, each transformer in the

population is reviewed individually by a set of
Subject Matter Experts (SMEs). We are interested
in the transformer as an asset subsystems or subassets, such as bushings, LTC selector or diverter
components, cooling equipment etc are considered
as part of the asset and if they must be maintained,
replaced or addressed in some way, that is considered
to be part of the normal transformer maintenance
and operation.
Many transformer owner/operators have a database

of laboratory DGA which can be used to give an
indication of individual transformer health, but also
rank transformers. Perhaps the easiest approach to
begin with is through application of an international
standard say IEEE C57.104 Guide for the
Interpretation of Gases Generated in Oil Immersed
Transformers (1).
325

April 8-10
2015

Within C57.104 there is a table of condition codes relating to both individual gases and total dissolved
combustible gas (TDCG), as shown below.
Table 1: Dissolved Gas Concentrations
The analysis suggested by C57.104 can be easily

implemented in a spreadsheet, or a common database
tool. The result is a population which is divided into
4 condition codes. However, the codes only indicate
degrading deterioration there is no implication for
what action may be taken, or how urgently it must
be done.
2.3 Combining Scores from Different

Condition Index Approaches
If we look at the different approaches to generating
a condition code, we can foresee problems where a
DGA is on a 1-4 base, and a Delphic approach is on
a 1-10. Each individual code should have similar

enumeration to other codes so that scores can be
combined putting everything in a 1-10 format
allows for calibration of analyses so that all code 6
have the same sense of urgency.
Combining via weights is a popular approach each
component score is weighted with a multiplier and
the final total score evaluated from all contributing
component scores. An example of combining scores
is given in the table below, with each parameter
weighted equally, with results for three different
transformers.
Factor
Scale
Transformer 1
Transformer 2
Transformer 3
DGA Main Tank
0-10
10
Dielectric
0-10
10
Thermal
0-10
Mechanical
0-10
10
Oil
0-10
DGA LTC Tank
0-10
Operational
0-10
10
Design/manufacturer
0-10
10
Subject Matter Expert
0-10
10
Sum
0-90
33
33
33
326

April 8-10
2015

It should be apparent in the above table that we have three transformers with identical overall scores. The
action required in each case is very different as the contributory factors to the score are different in each
case. We can say, however, that each transformer should have a similar urgency.
When we add weighting to the factors we can lose sight of the contributory factors. The table below shows
a system for weighting contributory factors for a tap changer.
Factor
Raw Score
Weight,%
Weighted Score
LTC Oil Condition
0.1
LTC Change in Oil Condition
10
0.3
Ethylene
15
1.2
Methane
15
0.3
Acetylene
0.1
Ethylene/Acetylene Ratio
15
0.6
LTC vs Main Tank Temperature differential
15
0.3
Days Since Pass Neutral, with 100 days being a 10
10
0.1
Operation Count
10
0.1
SUM
25
100
3.1
The result is a score but if the score changes from

3.1 to 3.5 what does that mean? Did the oil condition
change a lot or the methane change a little?
Logarithmic approaches to developing an index
based on contributory scores are used to retain the
indication of the root scores. The DGA score may be
assigned as 1 is good, and 100 is almost failed, but
the interim scores are only allowed to be: 3, 10 or
30. Any score which is 100 or greater requires
immediate attention as any one of the contributing
factors may be in an extreme state. We have thus
applied an urgency to the combined scores, and
cannot weight out the contributory factors.
2.4 Sanity Check

Once we have a score assigned for each transformer
we must now perform a sanity check that the
number of assets in each class/group is reasonable.
That is, we have an expectation of a failure rate based
on historic and industrial statistics which are
available (or not, as the case may be). Do we have a
similar number of condition based replacement
within the class/group assignments we have made?
If not, we should adjust the scores until we do have
something reasonable or something which is
different but which can be justified. We may have a
manufacturer or design group with known failure
mode which is different to the historic failure mode.
3.
CONCLUSION
In developing and applying asset health indices we

must make sure that the index reflects both the
condition and urgency for action. That is a useful
index as we can plan an intervention: replacement,
maintenance.
The index must be sanity checked in terms of:
-
an individual asset what the score means and

what an appropriate intervention plan is
the population and the number likely needing

replacement in short and long terms
It must be noted that it is possible to put together an

asset health index which is complex, comprehensive
and misleading especially if it has confusing weights
and scales. We must, therefore, be able to audit the
system to identify how scores are developed, what
they mean, and then justify the analyses used. It
can be easy to generate an index, but far more
difficult to justify its application!
BIBLIOGRAPHY
1.
327
IEEE C57.104 Guide for the Interpretation of

Gases Generated in Oil-Immersed Transformers
2008

April 8-10
2015

One Nation One Grid One Frequency

S.R. Narasimhan1
N. Nallarasan
Mohit Joshi
POSOCO, India
SUMMARY
The benefits of a large interconnected electricity
grid are reaped by all. Though the planning of
power system in India started with the concept of
regional self-sufficiency, the same was reviewed
considering resource concentration, load growth
areas and the importance of integration of
regional grids. This is furthermore important in a
country like India which has not only diverse
culture but also diverse power system portfolio
of each region. This diversity could be harnessed
to provide power to all at all-time only by
integrating the grids. In 1991, North Eastern
Region (NER) was synchronised with Eastern
Region (ER) followed by synchronisation of
Western Region (WR) in 2003 and Northern
Region (NR) in 2006. Till 2013, two grids namely
North-Eastern-North Eastern-Western (NEW) grid
and Southern Region (SR) Grid were in existence.
The synchronisation of SR with NEW grid was
done on 31st December 2013 over 765 kV RaichurSholapur. Prior to synchronisation, a number of
activities such as designing of additional System
Protection Schemes (SPS), testing of HVDC
frequency controllers and review of the Under
Frequency Relay (UFR) defense plans etc. were
done to ensure smooth synchronisation and
operation of National Grid. Oscillations of the
order of 0.2 Hz frequency was observed while
synchronizing and during other instances of major
load or generation tripping also. Subsequently,
in June 2014, the second circuit of 765 kV RaichurSholapur has also been commissioned which has
strengthened the AC interconnection between
NEW and SR Grid. Synchronisation of SR grid with
rest of the country has not only improved the
power transfer towards the southern region but
also brought about several improvements. The
1
Pradeep Reddy
frequency fluctuations have also reduced. The

success would pave the way for interconnecting
other countries and formation of a SAARC Grid in
future.
Keywords: Contingencies, Synchronisation,
Oscillations, Renewable
1.
BACKGROUND
Electricity grid in India has evolved from small state

system in 1960s to regional grids in 1980s and finally
to a National grid in early 21st century. Presently
there are five electrical regions in the country namely
Northern, Eastern, Western, North Eastern and
southern regions. Except southern region all the other
regions were integrated and known as NEW (North
East West North East) Grid. The important milestones
of the inter connection of the regions are as under.
October 1991: East and North East Indian grids
synchronized.
2nd March 2003: Western India Grid synchronized
with the East-North East Grid over 400 kV RaipurRourkela D/C line with 40% Fixed Series Capacitors
(FSCs) and +5-15% Thyristor Controlled Series
Capacitor or TCSC at Raipur to form the Central grid.
26th August 2006: Northern India Grid synchronized
with Central Grid over 400 kV MuzaffarpurGorakhpur D/C line with 40% FSC and 5-15% TCSC
at Gorakhpur end to form the NEW or North-EastNorthEast-West Grid.
Southern Grid was operating asynchronously with
the NEW grid but it was strongly connected with
the following HVDC links.
(i)
srnarasimhan@posoco.in
328
+/-500 kV, 2 x 1000 MW HVDC bipole from

Talcher in Eastern Region to Kolar in Southern
Region with 25% overload capability for a
duration depending on the ambient temperature
and cooling.

April 8-10
2015

(ii) 2 x 500 MW HVDC back-to-back station at

Gazuwaka in Southern Region.
recommended the regions to implement UFR

based load shedding, starting from 49.2 Hz and
going down to 48.6 Hz. Post synchronisation
of Southern Region with NEW grid, the stiffness
constant of the synchronous national grid would
have increased which requires larger load relief
from the constituents for the same amount of
increase in frequency. The revised UFR settings
as given below were to be implemented by
states prior to the synchronization of NEW grid
and SR grid:
(iii) 2 x 500 MW HVDC back-to-back station at

Bhadrawati in Western Region.
Initially, Southern Region was planned to be
synchronised with the NEW Grid over the 765 kV
Raichur (SR) - Sholapur (WR) 2 x S/C lines, with one
circuit being developed by POWERGRID and the
other circuit being developed by an Independent
Private Transmission Company, M/s RaichurSholapur Transmission Company Limited (RSTCL)
awarded through competitive bidding. These lines
were planned in 2007 as a part of Krishnapattanam
UMPP evacuation system for evacuation of power
from Southern Region to the other regions of the
country. Since, Krishnapattanam UMPP was getting
delayed and Southern Region was in acute power
shortage, it was decided in the planning horizon to
synchronise Southern Region with NEW grid over
765 kV Raichur-Sholapur 2*S/C irrespective of
Krishnapattanam UMPP commissioning for importing
power in Southern Region. 765 kV Raichur-Sholapur
line which was being constructed by RSTCL was
getting delayed. Considering the overall scenario and
the commissioning of 765 kV Sholapur-Raichur line1 by POWERGRID, it was finally decided to operate
Southern Region in synchronism with NEW Grid
through one 765 kV single circuit.
Table 1: Revised UFR Settings

Frequency
(Hz)
PRE REQUISITES FOR

SYNCHRONISATION
Considering that the synchronization entails

connection of a 85-90 GW NEW Grid with the 3035 GW Southern Grid, a number of activities were
identified which were to be completed before
synchronisation. Some of these activities are listed
below:
A.
NR
WR
ER
NER
Total
49.2
2160
2060
820
100
5140
49.0
2170
2070
830
100
5170
48.8
2190
2080
830
100
5200
48.6
2200
2100
840
100
5240
Total
8720
8310
3320
400
20750
B.
Installation of Phasor Measurement Units

(PMUs): Visualisation of the power system in
real time is not possible with traditional SCADA
which has inherent lags and delays.
Measurements from PMUs are generally taken
at high speed and help the system operator to
capture events which happens in milliseconds
which is useful for improving visualization and
situational awareness and taking necessary
actions. This becomes further more important
during smooth synchronisation of two large
grids. Therefore, it was decided to install PMUs
at Raichur and Sholapur and make available the
data to NLDC, SRLDC and WRLDC.
C.
Installation of interface meters at Sholapur

(PG) and Raichur (PG): Interface meters were
to be installed by CTU coinciding with
commissioning of Raichur and Sholapur
Substations as it is an inter-regional exchange
point to be used for energy accounting.
D.
Study of critical contingencies: In addition to

preliminary synchronization studies
(connecting through 765kV Raichur-Sholapur
D/C with floating), contingency studies were
performed; a number of credible contingencies
Anticipating synchronization of Southern Region

with the NEW grid, the National Load Despatch
Centre (NLDC) was estimating the quantum of power
flow on the NEW grid to SR AC link considering the
frequency difference, inertia and size of systems. It
was estimated that simply due to the diversity in
the two systems, power flow up to 800 MW in both
directions would take place.
2.
Regional Load Relief

in NEW Grid (MW)
Revision in Under Frequency Relay (UFR)

Settings: National Power Committee (NPC) had
329

April 8-10
2015

(iii) The SPS for contingencies in 400 kV

Raipur-Wardha-Parli-Sholapur section in
Western region was implemented.
were identified. Studies for the identified

scenarios were subsequently carried out. The
dynamic studies were carried out by applying a
three phase fault on 765kV Raipur Sholapur
one circuit and clearing the fault in 100ms by
tripping the circuit. Apart from credible
contingencies, impact of following low
probability high impact events were also
studied:
(i)
Outage of 2000 MW Kudankulam power

station without operation of SPS.
(iv) The SPS for contingencies of 400 kV

Hiriyur-Neelamangla and 400 kV GootyNeelamangla or 400 kV Gooty-Somanahalli
was implemented.
The design of SPS schemes is discussed in
detail in the subsequent sections.
F.
HVDC Frequency Controller testing: The

frequency controllers of HVDC Bhadrawati,
HVDC Gazuwaka and the Talcher-Kolar HVDC
were tested on 22nd Nov 2013, 19th Dec 2013
and 13th Dec 2013 respectively. These would
be put in service as per the settings elaborated
in section below.
G.
Primary Response from Generators: Free

Governor Mode of Operation (FGMO) /
Restricted Governor Mode of Operation
(RGMO) of generators is essential for operation
of a large synchronous grid and handle
contingencies. Accordingly, all the generators
were instructed to follow the provisions of IEGC
in this regard.
H.
Economy Interchange on AC Interconnection:

Based on discussion at CEA level, it was decided
that during the first six months of operation,
the 765 kV Sholapur-Raichur link would be used
only for reliability and economy interchanges
between the regions would be permitted only
after the experience of synchronous operation
during these six months and further
strengthening of upstream and downstream
network in the NEW grid and SR grid. Elsewhere
when the Turkish power system was
synchronized with continental Europe, an
extensive trail operation schedule was drawn
up with conomy interchanges coming much
later.
3.
FREQUENCY CONTROLLERS
SETTINGS
(ii) Tripping of Talcher-Kolar HVDC Bipole.

(iii) Tripping of Bhadrawati HVDC back to back
station carrying 1000 MW in West to South
direction.
(iv) Wind farm in Southern Region going off
the grid leading to 1500 MW generation
loss
(v) Outage of a large power station 1000 MW
and above in Southern Region.
(vi) Load crash in Northern Region leading to a
large surplus in the All India grid coupled
with heavy deficit in Southern grid.
(vii) Outage of a large power station in
Chhattisgarh zone leading to high voltage
in the system.
(viii) An incident at a large sub-station such as
Rourkela, Raipur, Wardha, Bhadrawati,
Parli etc.
E.
Installation and Modification of System

Protection Schemes (SPS)
Based on the results of the studies conducted
following SPS schemes were installed /modified
prior to synchronisation of Southern Region:
(i)
The SPS for 765 kV Raichur-Sholapur Ckt

1 line was commissioned for load shedding
in Southern region and generation back
down in Western region.
(ii) The SPS 1000 scheme of HVDC TalcherKolar Bipole was to be kept in service. In
addition to it, 600 MW additional
generation backing down in eastern region
as per the agreed scheme was implemented.
Modification to Kundankulam SPS was
implemented.
Automatic Frequency Controller is a device which

increases/decreases the power flow through the
HVDC link depending on the grid frequency
conditions and the target frequency. There are 3
HVDC links between NEW grid and SR grid i.e.,
330

April 8-10
2015

HVDC Talcher-Kolar Bipole, HVDC Gazuwaka BTB

and HVDC Bhadrawati BTB. All the three HVDC links
have inbuilt frequency controllers in them which
were tested before synchronisation of Southern
Region with NEW Grid. After commissioning of 765
kV Sholapur-Raichur section and synchronization of
Southern Region, the above HVDCs would essentially
become embedded HVDC. It became important for
the system operator to set the HVDC power order
optimally and retain control on the same. In case
frequency controller comes into operation under
normal conditions, the system operator loses control
on the power order. Therefore, the settings of the
frequency controller had to be carefully chosen so
that it comes into operation only during emergencies.
The frequency controllers normally can be kept ON
with the objective function defined for one side
(either NEW grid side or SR side) logically.
Considering that the Southern Region is the smaller
system as compared to NEW grid, the frequency
controller was activated for Southern Region side
objective function under abnormal conditions. The
transmission line loadings on either side also have
to be checked if the frequency controller has to be
activated. Considering that both the Talcher-Kolar
HVDC bipole and Bhadrawati HVDC are loaded to
2500 MW and 1000 MW respectively
simultaneously without any problem, the frequency
controller of these HVDCs on the Southern Region
side was activated. At HVDC Gazuwaka, the power
order is generally kept in the range of 600-650 MW
due to constraints in Eastern Region. Therefore, the
frequency controller on HVDC Gazuwaka was kept
OFF.
The above settings were based on the assumption
that the threshold of 48.8 Hz and 51.2 Hz on SR
side indicates that there is a separation of system.
This may not be entirely full proof. Hence, a more

scientific design based on sensing system separation
is required to activate the frequency controller.
4.
DESIGN OF SPECIAL PROTECTION

SCHEMES
The synchronisation of Southern Region was being

done over 765 kV Raichur-Sholapur S/C. Post
synchronisation, two large grid would be connected
through a single weak tie line. In such scenario, the
tie line would be carrying all the surplus and deficit
of the Southern Region. This would have meant
uncontrolled power flow on this tie leading to high
loadings on the other lines. Tripping of this tie line
under heavily stressed situation could have resulted
in cascading grid failure. In order to handle this
situation, a number of SPS schemes were designed
and some of the existing schemes were modified by
the system operator in consultation with the planners
till the time the interconnection is strengthened.
These SPS schemes are tabulated (Table 3).
The SPS have undergone changes with
commissioning of second circuit of 765 kV RaichurSholapur and other lines in Western and Southern
Region.
5.
SELECTION OF LAUNCH WINDOW
The selection of launch window for synchronisation

of Southern Region was dependent upon the
frequency profile of NEW Grid / SR Grid and voltage
profile of 400 kV Raichur and 400 kV Sholapur
substation. It is desirable that the frequency
difference between the two grids is not large so as to
avoid sudden inrush of power. Further, the voltage
at the two ends should be within the permissible
range.
Table 2: Frequency Controller Settings

Frequency Controller Settings
Setting
Gain Setting
HVDC Talcher-Kolar
HVDC
Gazuwaka BTB
Talcher End : 2000 MW/HZ

Kolar End : 1000 MW/HZ
HVDC
Bhadrawati BTB
Bhadrawati : 500 MW/Hz
KEPT OFF
Dead Band (NEW Grid Side)
48.5 Hz 51.5 Hz
48.5 Hz 51.5 Hz
Dead Band (SR Grid Side)
48.8 Hz 51.2 Hz
48.8 Hz 51.2 Hz
331

April 8-10
2015

Table 3: New SPS Schemes designed for synchronisation of Southern Region

Sl No
Initiating Condition
Action
Total flow on 765 kV Sholapur-Raichur 2 x 765 kV S/

C lines crossing a preset value, say 800 MW in
Sholapur to Raichur direction sensed both at Sholapur
and Raichur.
Tripping of 500 MW load in Southern Region
Rate of change of flow crossing 400 MW/second and

the final flow crossing in the Sholapur to Raichur
direction.
Loss of import on 765 kV Sholapur-Raichur by more

than 800 MW due to tripping of these lines.
Flow on 400 kV Raipur-Wardha any circuit crossing

800 MW sensed at Raipur OR tripping of any one
circuit of 400 kV Raipur-Wardha carrying 650 MW or
above.
800 MW Generating backing down in pre

identified plants in Western Region and 800 MW
Load shedding in Southern Region
Flow on 400 kV Wardha-Parli any circuit crossing 750

MW or tripping of any one circuit of 400 kV WardhaParli section carrying 600 MW or above.
800 MW Generating backing down in pre

identified plants in Western Region and 800 MW
Load shedding in Southern Region.
Flow on 400 kV Parli-Sholapur crossing 800 MW or

tripping of any one circuit of Parli-Sholapur carrying
600 MW or above.
Activates the 765 kV Sholapur-Raichur SPS

mentioned above and trips 500 MW load in
Southern Region
Tripping of one of the 400kV Hiriyur- Neelamangala

DC line and Flow on other line > 850 MW
Tripping of 500 MW loads in Southern Region

and backing down of 500 MW generation in
Western Region
Tripping of one of the 400kV Gooty-Nelamangala SC

line or 400kV Gooty- Somanahalli SC line and Flow
on other line > 850 MW
Tripping of 500 MW loads in Southern Region

and backing down of 500 MW generation in
Western Region
Fig. 1: Frequency difference between NEW and SR Grid from

16/12/2013 till 22/12/2013
Fig. 2: Voltage Profile of 400 kV Raichur from 16/12/2013 till

22/12/2013
It can be inferred from the above graphs that while

the frequency difference is relatively small and steady
during the night hours, the Sholapur and Raichur
voltages are generally high during this period.
Therefore, the launch period for synchronisation of

NEW and SR Grid was selected from 1700 hrs to
2100 hrs keeping in view the availability of hydro
generation during this time.
332

April 8-10
2015

(vi) Power Oscillation Damping (POD) feature of all

HVDCs and TCSCs were kept in service.
(vii) All short term transactions towards SR were
curtailed initially from the day of SR Grid
synchronisation with NEW Grid for intial few
days so as to ensure adequate margins on the
HVDC back to back stations and bipole links
connecting to Southern Region to control
loading on 765 kV Raichur-Sholapur.
(viii) The maximum continuous power order on
HVDC Talcher-Kolar Bipole was restricted to its
maximum continuous loading of 2000 MW. The
500 MW margin for 10 hrs per day was kept to
handle operational exigencies.
Fig. 3: Voltage Profile of 400 kV Sholapur from 16/12/2013 till

22/12/2013
6.
SYNCHRONISATION OF SOUTHERN
REGION
The synchronisation of Southern Region was planned

on 31st December 2013. Following were ensured
before synchronisation:
(i)
All transmission elements in major corridors

were healthy and in service.
(ii) Frequency, Voltage, Line Loadings and Phase

Angle difference between adjacent buses were
in operating range.
(iii) All protection systems, SPS and Islanding
schemes were in place and kept operational.
(iv) All the entities are within generating / drawing
as per the schedule.
(v) Black-start facilities wherever available were
kept ready. System restoration procedure,
Power-maps and other logistics were also kept
ready in all control centres.
Prior to synchronisation, the angular difference

between Sholapur and Raichur was within +/-5
degree and the frequency difference between SR and
NEW grid was brought down to below 0.05 Hz by
following actions.
(i)
Changing Power Order of Bhadrawati HVDC,

HVDC Talcher-Kolar and HVDC Gazuwaka
(ii) Variation of Hydro Generation in Southern

Region.
(iii) Variation of Thermal Generation to the extent
possible.
(iv) Load Management in NEW Grid and SR, if
required
SR Grid frequency was kept higher than the NEW
Grid frequency. The line was first charged from both
the ends to check the healthiness of the line.
Thereafter, the line was first charged from 765kV
Raichur end. Both grid frequencies were adjusted as
Fig. 4: Mimic of Synchroscope implemented at NLDC through PMU data
333

April 8-10
2015

Fig. 5: PMU plots of Frequency, Voltage magnitude of Raichur/Sholapur and 765kV Raichur-Sholapur line flow
per the requirement mentioned above. Operator at

Sholapur was kept over telephone. A mimic of
synchroscope was prepared at NLDC using PMU
measurements which aided the operator in giving
closing instruction to the operator at Sholapur substation. Finally the synchronisation was done at 2025
hrs. The oscillations of the order of 0.2 Hz frequency
has been observed for 15-20 seconds while
synchronizing the SR grid with NEW grid.
7.
OPERATIONAL EXPERIENCE
1.
Initially, on two (2) occasions, the line tripped

when the power flow on the line crossed 800
amperes while on the other two (2) occasions,
the line tripped due to tripping of the 2 x 1500
MVA 765/400 kV Inter Connecting Transformers
(ICTs) at Raichur reportedly due to DC earth
fault. The power flow on 765 kV SholapurRaichur varied up to 800 MW in either
directions intially, particularly at the hour
boundary when there was a sudden change in
load. The loading of 765kV Raichur-Sholapur
line was controlled by changing the set-points
of Bhadrawati and Gazuwaka back-to-back
HVDC and Talcher-Kolar HVDC Bipole. On a
net energy exchange basis, the power transfer
over the line over a day was negligible.
Fig. 6: Power flow variation on 765 kV Raichur-Sholapur line
Fig. 7: Energy transfered to Southern Region in 2014
2.
The pattern of 765kV Raichur-Sholapur line

flow during the first month of synchronisation
and the power transfered to Southern Region in
2014 is given below :
334
The oscillations of the order of 0.2 Hz frequency

was observed for 15-20 seconds while
synchronizing the SR grid with NEW grid on all
the occasions till date. The PMU plots of
oscillations are given below :

April 8-10
2015

Fig. 8: Oscillations during synchronization of 765 kV Raichur-Sholapur-I at 1041hrs on 05.01.14
3.
The performance of System Protection Schemes

(SPS) has been mixed. On some occasions, less
response was observed from Special Protection
Schemes (SPS) due to various reasons.
4.
Apart from the above oscillations at the time of

synchronization, spontaneous oscillations of
frequency of the order of 0.2 Hz have also been
observed on other occasions also. In normal
operation, the 0.2 Hz oscillations were getting
damped out within 20-25 seconds. However

there was one incident when the oscillations
persisted for two (2) minutes on 28th Jan 2014.
Another aspect of these oscillations was that
the Agra bus in Northern Region swings in phase
opposition to Sholapur/Raichur which suggested
that there was a problem of both synchronizing
power as well as damping. The PMU plot in
one such event is given below:
Fig. 9: PMU Plot of 765 kV Raichur-Sholapur line flow and frequency of 765 kV Agra & 765 kV Sholapur at 1620 hrs on 22.01.14
335

April 8-10
2015

Fig. 10: Plot of Maximum, Minimum and Average frequency of SR from 2013 to 2014
5.
6.
Frequency fluctuations have gradually reduced

since synchronisation of Southern Region with
NEW Grid.
It has been observed that due availability of
more power during the day, Kadamparai
Pumped storage hydro plant has done more
pumping of water. The energy consumed by
Kadamparai pumped storage hydro plant has
increased from 309 Mus in 2013 to 581 Mus in
2014. This has helped the system operators to
better handle renewable energy in the system.
8.
CONCLUSION
Formation of National Grid has been one of the

greatest achievement of this decade. Utmost care
was taken in the entire process and emphasis was
placed more on reliability considering the 765 kV
Sholapur-Raichur line as a hand-holding connection
only rather than a connection on which one could
lean on heavily. Though at present only limited
amount of power is allowed to flow over the AC
interconnection between NEW and SR Grid, the
power transfer to southern region is expected to
Fig. 11: Comparison of energy consumed by Kadamparai in Pumping mode
336

April 8-10
2015

improve with the commisioning of other important

lines in Western and Southern Region. Other
important steps such as tightening of frequency band,
implemntation of primary and secondary control,
development of parallel corridors etc are required to
reap full benfits of national grid.
Acknowledgement
Authors are grateful to the power system fraternity
and POSOCO Management for the encouragement.
Authors also express their gratitude to all those who
have contributed towards synchronisation of
Southern Region. The views expressed in this paper
are those of the authors and not necessarily of the
organization they belong to.
BIBLIOGRAPHY
1.
available on https://sustainabledevelopment.
un.org/content/documents/interconnections.
pdf
2.
NLDC Operational Feedback on NEW and SR

Grid Synchronisation dated 16th Jan 2014, 07th
Feb 2014 and 11th Apr 2014
3.
Central Electricity Regulatory Commission,

Indian Electricity Grid Code available at
www.cercind.gov.in
4.
Electrical Interconnection between Turkey and

Europe: Problems and Solutions, Prof. Dr.-Eng.
Harald Weber, 13th Middle East Power Systems
Conference, MEPCON2009, Assuit University,
Egypt, December 20-23, 2009
5.
Technical Background and Recommendations

for Defence Plans in the Continental Europe
Synchronous Area, ENTSOE, 2011
United Nations, Multi Dimensional Issues in

Electric Power Grid Inter Connections 2006
337

Gridtech Subs Auto Smart

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Gridtech Subs Auto Smart

Hochgeladen von

Copyright:

Verfügbare Formate

SESSION

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

Special Protection Scheme for Bus Reactor in

modality, there are challenges regarding (i) Bus

In recent years, power flow pattern has changed in

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

in non-agriculture period. Details are mentioned in

are required to be integrated precisely without

(i) Switch Yard Arrangement

Month wise Maximum Voltage (KV)

Bus Reactor is provided in front of Transfer Bus

Connection through Transfer Bus

Connection through bottom of Pantograph

In case of first option, issues would be as under:-

(a) Physical connectivity with Transfer Bus

(b) Bus Reactor will always remain connected

To over come this issue, 400KV Bus Reactor is

In order to consider Bus Reactor in TBC Bay,

(c) Transfer Bus will unnecessarily remain

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

through Bottom of TBC bay Transfer bus

Separate isolator with earth switch is to be

No separate CTs are to be provided for Bus

(iii) Control Philosophy

No separate control for Reactor Bay is

Annunciator in existing TBC Control panel

(iv) Interlocking Philosophy

Reactor Isolator cannot be closed until

Transfer Bus Isolator of TBC cannot be

(ii) Bay Philosophy

Bus Reactor is to be connected to TBC Bay

All bay Transfer Bus Isolator control

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

(v) Bus bar protection scheme Philosophy

TBC Bay CT core is routed to Z-3 of RADSS

Reactor Differential Protection: TBC Bay

Reactor Back-up Impedance Protection:

Reactor REF Protection:Reactor Bus side and Neutral side Bushing

(vi) CT Core utilization for Metering and

Reactor Bus bar Protection:No need of separate CT core as it is not a

Metering: Reactor Bushing CT core is

(Reactor Bushing CT cores towards Bus side

As this is not being a regular Reactor bay, scheme

Separate Control Panel for Bus Reactor is not

All Reactor protection and Reactor Trouble

Trip circuit supervision relay, CVT selection

Separate energy meter is required for Bus

5th International Exhibition & Conference, 2015

New Technologies in T & D, Renewable Energy Integration,

The above requirements are summarized in Table 2.

Reactor isolator is open.

It is necessary to prepare operation philosophy for

Bus Reactor has its own Isolator with E/S.

Remaining equipments of TBC bay shall be

TBC Panel scheme such as CB Close and Open,

When TBC Bay is to be utilized as Bus Reactor

Bus Reactor is connected to Transfer Bus

Separate Reactor protection panel is provided