21, rue d'Artois, F-75008 Paris

http://www.cigre.org

33-105

Session 2002
CIGR

INSULATION COORDINATION STUDIES FOR 400 kV GIS IN A HYDROELECTRIC

PROJECT IN INDIA
B.N. SARKAR*
K.S.MEERA

GAURAV GARG
RAM NATH
A.K.TRIPATHY

CENTRAL POWER RESEARCH

INSTITUTE
J.THOMAS

S.C.MISHRA
TEHRI HYDRO
DEVELOPMENT
CORPORATION LTD.

BHARAT HEAVY
ELECTRICALS LTD.

INDIAN INSTITUTE OF SCIENCE

INDIA
Abstract: Insulation coordination studies comprising of
Switching, Lightning and Very Fast Transient overvoltages
has been carried out for the 400 kV Gas Insulated
Switchgear (GIS) in the Tehri hydroelectric power plant in
India. The studies were carried out to decide the various
parameters of the switchyard equipment including arrester
locations. Results of the switching overvoltages have been
validated on the Transient Network Analyser (TNA). This
paper discusses some of the results of all these studies.

to Tehripool and two 765 kV single circuit lines from

Tehripool to the load centre Meerut. The single line
diagram of the scheme is shown in figure 1. System studies
were carried out to determine the overvoltages likely to
develop in the Tehri 400 kV GIS due to switching of any
of the 400 kV transmission lines, due to lightning and also
due to disconnector operations. The studies were carried
out using the Electro Magnetic Transients Program
(EMTP) and EMTDC software.

Key words: GIS Switching,- Lightning - Very Fast

Transient Overvoltage - Insulation Coordination
Modeling - Disconnector

2.0 SWITCHING OVERVOLTAGE

1.0 INTRODUCTION
Tehri Hydro Electric project is one of the major hydro
projects under execution by the Government of India. The
hydro station is situated in the foothills of the Himalayan
range. Due to geographical constraints the generator and
turbine hall, transformer hall have been located inside a
tunnel in the mountain. This project owned by Tehri Hydro
Development Corporation (THDC) is scheduled to be
commissioned in two stages. Stage I of the project would
comprise of four numbers of 250 MW hydro machines at
Tehri in an underground powerhouse. Stage II would
include additional 4*250 MW machines at Tehri and
4*100 MW machines at a surface powerhouse at
Koteshwar about 20 km downstream from Tehri.
A total power of 2400 MW is to be evacuated from the
generating station at Tehri to major load centers including
New Delhi. The evacuation scheme consists of Gas
Insulated Bus ducts (GIB) having a three-phase run of 800
m in an underground tunnel from Tehri, a triple circuit 400
kV line from Tehri to a power pooling station - Tehripool,
a double circuit 400 kV transmission line from Koteshwar

Switching overvoltage stressing the overhead lines are

originated by sudden changes in the system configuration,
either operation of a circuit breaker - line energisation or
reclosing, or occurrence of a fault. The three-phase
reclosing and line energisation operations, which generate
the maximum switching overvoltage (SOV), are
considered in the study. Also, the use of pre-insertion
resistors and surge arresters are the only measures
considered here for limiting of SOV.
For SOV studies, distributed line model was used for
representing the transmission lines. The GIS bus ducts
were modeled by lumped capacitances - phase to ground
and the synchronous generators as constant voltage
sources behind transient reactance. The generator and the
step-up transformers were represented by three-phase
models with correct vector group and saturation
characteristics. The zinc oxide arresters were modeled
with type 92 arrester model in EMTP. On TNA the
transmission lines were modeled by a series of pi sections
and the transformers by appropriately connected singlephase transformer models along with their saturation
characteristics. The SOV studies covered a number of
operating conditions comprising of varying number of
generators, bus ducts and
transmission
lines
to

Central Power Research Institute, Sadashivnagar P.O., P. B. No. 8066, Bangalore, Karnataka- 560 560 080, India

evaluate the maximum magnitude of the overvoltage.

For each case 200 statistical switching operations were
considered. A pole span of 0-2-4 ms and pre-insertion
resistors (PIR) of 400, 600 and 800 inserted for 8, 10
and 12 ms were considered. In case of line re-closing

2 * 100 MW

studies trapped charges were also simulated. On TNA,

selected cases from EMTP - those yielding the maximum
overvoltage were studied. The results of the EMTP and
TNA studies are presented in Table I. The overvoltage
indicated here are peak values.

TEHRIPOOL
400 kV

11/ 400 kV
2 km

G
220 MVA

G
/Y

2 * 100 MW

MEERUT

KOTESWAR
180 km
TEHRI

20 km
/Y

G
15.75 kV
(8 * 250 MW)

15.75/400 kV
315 MVA

800 m
TEHRI

189Mvar

TEHRIPOOL
Y/Y
3*1000 MVA
400/765 kV

Fig 1. Single line diagram of Tehri transmission system

3*1000 MVA
765 /400 kV

Table I: Maximum Switching Overvoltage

Switching operation

PIR & Insertion time

Maximum Overvoltage in p.u.

Line Energisation (Tehri 400 kV breaker)*

Line Energisation (Tehri Pool 765 kV Breaker)**

No
400 , 8 ms
No

EMTP
2.29
1.63
2.69

TNA
2.10
1.57
2.52

600 , 12 ms
No

1.74

1.65

3-Phase Line Reclosing (Tehripool 765 kV Breaker) **

3.79

3.68

600 , 12 ms

1.82

1.71

* 1p.u. =326.59 kV peak, ** 1p.u. = 624.62 kV peak

Figure 2a shows the energisation overvoltages at
Tehripool bus for switching from Tehri end, without PIR
as obtained using EMTP. Figure 2b shows the same as
obtained using TNA. Figure 3a shows the voltage
waveforms at Tehripool bus for switching at Tehripool,
with PIR of 400 ohms included for 8 ms. Figure 3b shows
the same as obtained on TNA.
2.1 Results
The results of the switching overvoltage studies indicate
that to limit the overvoltage arising due to energisation
and re-energisation in the 400 kV Tehri network to within
2.5 p.u., a PIR of 400 inserted for 8 ms is to be used at
Tehri 400 kV breaker. Similarly, a PIR of 600 inserted
for 12 ms is to be used in 765 kV breaker at Tehripool to
limit the overvoltages within 1.9 p.u. in the 765 kV
network. Thus a rated switching impulse withstand level
of 1050 kV has been recommended. As seen from the
oscillograms and the results of Table 1 the simulation
results of EMTP and TNA tally well with each other
thereby validating the correctness of the simulation.
3.0

LIGHTNING OVERVOLTAGE STUDIES

Lightning overvoltage studies were carried out using

EMTP. The lightning surge was represented by a
standard impulse of 1.2/50 s and was applied within
the 1st span of out going line from switchyard.
Number of out going lines from switchyard was seen
to have influence on lightning overvoltage (higher the
number of outlets, lower was the overvoltage). Hence
only one line and one transformer case was considered.
Open disconnectors or open bus bar were seen to give
highest overvoltage due to doubling effect. Different
combination of generator and outgoing transmission
line were considered. The overvoltage at any point
other than transformer was limited to 1140 kVp and
near transformer it was limited to 1083 kVp keeping
20% margin with the BIL level of 1425 kVp for
switchyard and 1000 kVp for transformer. In order to
adequately protect the transformer and GIS equipment,
it was necessary to provide surge arresters at following
locations:
i)
A GIS arrester close to every transformer

ii)
iii)

A GIS arrester on each of the outgoing GIB

feeder just after the disconnected switch
An open air arrester just after the SF6 to air
bushing on each outgoing line

With a rating of 390 kV, maximum energy rating of

arresters were within 1.5 kJ/kV. Maximum voltage at
400 kV buses could be arrested to 990 kVp and at
transformer terminals to 898 kVp.
4.0

VERY FAST TRANSIENT OVERVOLTAGE

The operation of disconnector switch and circuit

breaker in a GIS typically involves numerous restrikes.
These restrikes could generate a large number of Very
Fast Transient Overvoltage (VFTO) with a rise-time of
the order of a few ns followed by high frequency
oscillations. The peak value of the VFTO could be as
high as 2.5 p.u or more in extreme cases [1,2]. VFTO
can lead to Transient Enclosure Voltage (TEV) as well
as ElectroMagnetic Compatibility (EMC) related
problems to sensitive electronic circuits used in
metering and controlling equipment. VFTO can also
cause low probability flashover to ground especially
for system voltages above 400 kV. Hence it is
important that an estimate of the VFTO magnitude is
made before commissioning of substations.
In the present work, the VFTO levels were estimated
for different switching conditions and at various nodal
points of the proposed 400 kV GIS. A one-line
diagram of the 400 kV GIS for the Tehri Hydro project
is shown in Figure 4. This substation has 4 Stage I
generators (G1 to G4) and 4 Stage-II generators (GAl
to GA4) feeding the three buses and three outgoing
feeder lines from the substation.
4.1

Computer Simulation

By concept the GIS behaves like an interconnected

transmission line when subjected to fast transients.
Therefore, EMTP can be used for simulation of VFTO
in a GIS. For estimating VFTO level, simulation of all
the components of the GIS and other connected
equipment are required. The transient caused by a

single spark is damped in few s, i.e., much earlier

than a subsequent spark, if any occurs. For this reason
VFTO simulation can be conducted by considering
each strike separately. VFTO levels of a GIS
configuration depend on the observation point as well
as the switching location in addition to the impedance
of its various components. Since the VFTO contain
predominantly high frequency components ranging
from hundreds of kHz to tens of MHz, most of the
models have their capacitances dominating the other
parameters.
The typical lengths of a GIS bus are much smaller than
an ordinary substation and also, at high frequencies in
the range of few hundreds of kHz to MHz, the GIS bus
acts like a transmission line with a finite transit time
and propagation velocity. The value of surge
impedance of GIS bus bar which is modeled as
transmission line can be obtained from the relation Z =
60 ln (b/a) , Where a is the diameter of the HV bus
and b is the inner diameter of the enclosure and Z is
found to be 64.2 for the GIS bus.
Elbows, spacers and spherical shields are modeled by a
lumped capacitance of 15 pF. Power transformer with
bushing mounted is modeled as lumped capacitance to
ground of 1250 pF whereas surge arresters are
modeled as lumped capacitance to ground of 50 pF.
Disconnector switches are modeled differently under
open and closed condition. When opened they are
modeled as two transmission lines in series with a

capacitance. The surge impedance of each of the

transmission line is taken as 69 whereas the series
capacitance is chosen as 20 pF. A capacitance to
ground of 52 pF is included at both ends of the switch.
When closed, the capacitance in between the two
transmission lines of the open disconnector switch
model is replaced by a transmission line with the same
parameters as before but the capacitance to ground is
taken as 88 pF.
The spark is modeled as an exponentially decaying
resistance. Bushings are modelled with a lumped
capacitance to ground of 500 pF whereas earth
switches are modeled by lumped capacitance of 45 pF.
Circuit breakers are also modeled differently under
open and closed condition. The open circuit breaker is
modeled as two transmission lines in series with a
capacitance of 2800 pF and a 17 resistance in series
between them. The capacitance to ground is 200 pF.
The closed circuit breaker is modeled similar to the
open circuit breaker with the capacitance replaced by a
transmission line of surge impedance of 30 Ohm. The
capacitance to ground is replaced by 20 pF. The
overhead line, which is connected to the GIS, is
considered to be of infinite length. Hence it is modeled
as a transmission line terminated with a resistance to
ground of value equal to the surge impedance of the
overhead line (350 ), so that there is no reflection
from this end.

Fig 4 Single-line diagram of 400 kV GIS for Tehri Hydro Project

Figure 5 VFTO Waveforms

A total of 61 different switching conditions have been

simulated for a trapped charge magnitude of 1.0 p.u.
Computer simulations have been done with EMTP and
using appropriate transient models for the GIS
components.
4.2

Results and Discussion

Only two specific cases have been presented and

discussed here. The conditions considered are
a)

All generators connected and the disconnector

switch between M11 and M14 operated with
circuit breaker between M16 and M19 kept open.

b) All generators connected to the bus and the circuit

breaker between M16 and M19 operated.
4.3

VFTO waveforms

The computed VFTO waveforms for disconnector and

circuit breaker switching conditions are shown in
Figure 5. In general it is seen that when the observation
node is close to the switching point, the VFTO
waveform has higher frequency content as compared to
a farther away node from the switching point. Also, it

Figure 6: Variation of peak VFTO levels for outgoing feeders

along Nodes (Disconnector Operation)

is found that the VFTO peak magnitudes are higher in

the case of circuit breaker operation as compared to
that of disconnector switching operation. Out of all the
61 switching operations simulated, amongst
disconnector switching operations, VFTO peak is
found to be highest when the disconnector switch
between M226 and M229 is operated with the circuit
breaker between M26 and M29 kept open. Its value is
l.67 p.u. (node M26). In the case of circuit breaker
operation, the VFTO peak magnitude is found to be
maximum when circuit breaker between M16 and M19
is operated. Its value is 2.7 p.u. at the node M123 on
the outgoing feeder X=1.
4.4

Variation of VFTO peak along the nodes

The variation of the VFTO peak along the nodes of the

outgoing feeder lines for each of the above mentioned
switching conditions are shown in Figures 6 and 7.
There is a distinct pattern of variation of VFTO peak
along the nodes of the GIS in the case of disconnector
switch operation as compared to that of circuit breaker
operation (compare Fig 6 and Fig 7). From Figure 6 it
is seen that the outgoing feeder line Ml shows a peak
value close to 1.47 p.u. The VFTO magnitudes reduce
near the open circuit breaker and once again increase

Figure 7: Variation of peak VFTO levels for outgoing feeders

along Nodes (Circuit Breaker Operation)

near the junction of the GIS with the overhead

transmission line. From Figure 7, it is seen that the
outgoing feeder line M1 (X=1) has a maximum VFTO
value of 2.70 p.u. There is a steady increase in the
VFTO levels beyond the switching point till the
junction of GIS with overhead line in the case of
circuit breaker switching operation.
5.0

CONCLUSIONS

Switching overvoltage studies decided the magnitude

of preinsertion resistor and the duration of its insertion
to keep the switching surge overvoltage under control.
Lightning overvoltage studies revealed that additional
arresters have to be installed for proper insulation
coordination in the GIS station.
The VFTO peak magnitudes are higher close to the
switching point in the case of disconnector operations
where as they are higher at the junction of the GIS with
overhead line in the case of circuit breaker operation.
Also, close to the switching point, VFTO waveform
has higher frequency of oscillations as compared to an
observation point farther away. There is also a distinct

pattern of variation of VFTO peak along the nodes of

the GIS in the case of disconnector switch operation as
compared to that of circuit breaker operation. The
circuit breaker operation results in the highest VFTO
level amongst all the switching operations done. Based
on these studies, the rating, location and energy levels
of the arresters was finalized to evolve a good
protection philosophy for insulation coordination of a
400 kV GIS station.
6.0

REFERENCES

1. J. Lewis, B. M. Pryor, C. J. Jones, T. Irwin,

"Disconnector Operations in Gas Insulated Substations
Overvoltage Studies and Tests Associated with a 420
kV Installation", CIGRE, Vol. lI, 1988, paper 33.09,
pages 1-8.
2. Joint working group 33/23.12, "Insulation coordination of GIS: Return of experience on site tests
and diagnostic techniques", Electra, February 1998,
pages67-97.