Cтатья: Tеst Models for Explosion Protection of High Voltage Oil Filled Electrical Equipment | Журнал «ЭЛЕКТРОЭНЕРГИЯ. Передача и распределение»

Page 1

The MAIN JOURNAL for POWER GRID SPECIALISTS in RUSSIA

Page 2

August 25–29, France, Paris

Equipment Protection

Tеst Models for Explosion

Protection of HighVoltage

OilFilled Electrical Equipment

Leonid DARIAN (

Леонид

ДАРЬЯН

Vladimir POLISCHUK (

Владимир

ПОЛИЩУК

Aleksey SHURUPOV (

Алексей

ШУРУПОВ

CJSC “Technical inspection UES”,

Russian Federation Joint Institute for High Temperatures of RAS

INTRODUCTION

Life time of transformers or another HV oil-

lled

electrical equipment (HV OFEE) is about several decades.
The gradual degradation of paper-oil insulation occurs
under the in

uence of partial discharge, heating, cavitation

and other factors in service. In time the deterioration of
insulation characteristics exceed a critical level, that’s
why untimely out of service may cause arc discharge
(AD) due to internal short circuit (ISC). Electric power
of discharge may range from tens to hundreds MW. Large
amount of hydrogen, oxides of carbon and hydrocarbon
gases is formed due to decomposition of transformer oil
(TO) under action of AD. Due to TO incompressibility,
formation of gases causes the raise of pressure that quite
often ends by explosion of HV OFEE body. Mixture of
atmospheric air and hot hydrogen, oxides of carbon and
hydrocarbon gases can ignite the in

ammation. In this

case the damage from explosion increases many times,
and the work of electrical power substation is suppressed
for a long time. According to data from foreign and
native sources, the possibility of early out of service is
10

-2

, means every 1 from 100 transformers will be out of

service due to explosion. The possibility of in

ammation

after HV OFEE explosion is about 15%.

In case of severe accidents the losses determined

by cost of the replacing equipment can amount to tens
of millions dollars. Therefore, the improvement of
explosion protection for HV OFEE is important for
the electrical power industry. As time goes, the lack of
appropriate technical and organizational solutions will

make this problem worse. Firstly, there is a general trend
of increasing equipment capacity, and secondly, it’s
not always possible to provide an adequate renewal of
equipment.

The necessity of explosion protection improvement

for HV OFEE is coming from the project "Basic
principles (Concept) for Technical policy of Russian
electric power industry for the period up to 2030" and the
"Regulations of technical policy for JSC "FGC of UES".
According to HV OFEE operating conditions, we cannot
completely exclude the possibility of internal short circuit.
Nevertheless we can achieve reduction of accident risks as
well as loss reduction by improving HV OFEE design and
its protection systems.

The destruction degree of HV OFEE is mainly

determined by energy

released in AD. The energy

depends on AD duration (or action time of protection
devices), point of ISC origin, characteristics of external
circuit. According to the literature data range of

possible values for industrial HV OFEE exceed by two
orders of magnitude. For example,

xed

energy values

in 735 kV power transformers are in a range of 1 to 147
MJ. [4] 735 kV single-phase transformer tank exploded
at AD energy level of more than 8 MJ, but

re started at

energy level of more than 14 MJ. However, even at

energy level of more than 20 MJ not every transformer
explosion was accompanied by

re.

energy level of 330 kV instrument transformers is

about 0.3—1 MJ, Qa energy level of 100 MVA distribution
transformers can be varied in the range of 3—10 MJ.

Page 3

[email protected], www.eepr.ru

energy level of more powerful transformers and boxes of
bushings under AD can be tens of MJ.

Loss reduction can be achieved in several ways

including:
• improvement of diagnostic techniques for degradation

level of TO and POI insulating characteristics in order
to take an equipment out of service for repair and
maintenance activities;

• reduction of protection devices action time;
• using of alternative insulating

uids with improved

performance instead of mineral transformer oils;

• creation of a non-explosive HV OFEE design and

improvement of protection systems.
In response to this problem needed an effective test

model for equipment under the impact of high pressure
pulse which occurs in AD. The standard test model for
explosion protection of HV OFEE based on method of
electric arc ignition in the internal volume of HV OFEE.

Methods [5—9] present the results of researches which

justi

ed an alternative test model for explosion protection

of HV OFEE. In this method a high pressure pulse which
occurs after ISC was simulated by chemical energy of
explosive materials (EM). The new method allows us to
carry out tests directly on place of manufacture or on site
of HV OFEE instead of expensive test facilities. Figures
show that the alternative tests will cost much cheaper than
standard tests, besides this method would be cheaper for
larger HV OFEE.

Joint Institute

High Temperatures of RAS as per JSC

“FGC UES” order developed an arcless source of pulse
pressure (ASPP) for explosion protection tests of HV
OFEE. Tests of protection system prototypes were carried
out in HV OFEE models at energy effect up to 5 MJ by
now.

In this study we present test results of protection system

prototypes for HV OFEE. This work also summarizes
the research results of AD in transformer oil, which was
the basis for developed ASPP, and application of this
method for explosion prevention tests of serial instrument
transformers 110 kV/330 kV.

According to accepted de

nition of explosion-proof

electrical equipment — it is electrical equipment, where
may occur structural damage under ISC, however all its
components must be inside the normable safety area close
to the equipment, which is calculated as the diameter
(width) of the equipment increased by two of its height
but not less than 1.8 m.

ARC DISCHARGE IN TRANSFORMER OIL

Our research results of AD in test model of HV

OFEE are described in detail in studies[5—8]. The AD
basic parameters resulting from these studies are below.
Experiments [5—8] were carried out under conditions
which are similar to conditions after ISC occurrence in
industrial HV OFEE, where discharge current increases
up to 10—30 kA in 3—10 ms. In our experiments the
critical discharge current reached 30 kA at the rise time of

1—3 ms. The total discharge duration was 3—20 ms. The
maximum heat release in the arc

reached 0.1 MJ. The

capacitive storage acted as energy source with maximum
voltage of 5 kV. Used electric circuit allows to simulate
two half-wave of heteropolar current, however basic
experiments were carried out with one half-wave voltage.
The maximum of AD average power is reached during the

rst half period, hereafter discharge voltage and power are

reducing due to resistivity degradation of insulating

uid.

AD was ignited between two parallel brass electrodes

about 20 mm in diameter. The distance between electrodes
was varied from 17 to 30 mm. The electrodes were located
in a body of 310 mm internal diameter and of 61 liters
volume. The liquid volume was 35 liters. The remaining
volume (26 liters) was

lled with nitrogen at atmospheric

pressure. The electrodes were close to body axis. The
distance from point of discharge origin to “liquid-
nitrogen” interface was 100 mm.

AD current and voltage, the pressure in the body, the

pressure of gas bubble above liquid were measured during
the experiments. Response time of pressure sensor was
less than 0.5 ms. One pressure sensor (PS) was installed
close to body foot (PS1), the other was installed at
50 mm from upper level of the liquid (PS2). We used high-
speed shooting of the discharge with a time resolution of
0.1 ms and point of discharge origin of “liquid-nitrogen”
interface with a resolution better than 0.8 ms. Shooting
was carried out through a side window of 150 mm
diameter. The amount of hydrocarbon gases formed due
to decomposition of transformer oil (TO) was calculated
from pressure value.

The experiments were carried out using different

liquids: mineral transformer oil of GC brand, vegetable
oils and distilled water. The research data of AD in TO is
below.

The arc discharge was initiated by applying voltage

(~ 3 kV) to a copper wire of 0.1 mm diameter connecting
electrodes. Fig. 1 shows oscillograms of current and
voltage in AD. The discharge duration (~ 7,5 ms) is close
to half-wave voltage duration at power frequency. At start
time the voltage oscillogram is on a sharp rise, and then
there is a rapid decline. Hereafter the discharge voltage
was varied in a small range. The estimated high voltage
peak duration (~ 20 ms) is equal to electric explosion time
of copper initiator.

There are some voltage pulsations at current fall time,

which are probably associated with arc movement on
the surface of electrodes. Arc movement rate is about
20 m/sec. Analysis revealed that AD column extends
under action of its magnetic

eld with the result of AD

voltage increase which causes shunt breakup with further
voltage decrease. Figures show that typical electric

eld

value in AD column is 0.1—0.3 kV / sm.

High-speed shooting of the discharge showed that

plasma glow was concentrated close to electrodes at the
beginning. At this point the glow area started to expand at
a rate of 0.3 km/s, however in 0.5 ms the rate decreased

Page 4

August 25–29, France, Paris

Fig. 1. Current and voltage oscillograms (current vs. time)

approximately threefold. Thus, the
plasma expansion rate was much
lower than the speed of sound in
TO, which is about 1.4 km/s [10].
Radiation

ashed over the electrode

spacing in about 1 ms after AD
occurrence.

Fig. 2 and 3 show liquid pressure

"oscillograms" from the same
experiment as current and voltage
oscillograms in Fig. 1. As Fig. 2
and Fig. 3 show that variation of
pressure in TO is repetitively-pulsed.
It could be seen quite distinctly
especially in the beginning of arching
which is about 3 ms. The

rst six

pressure extremums (maximum and
minimum) were registered with PS1
and PS2. The time difference between
them were approximately 0.1 ms.
This time delay was the same as time
difference of sound propagation in
TO from AD burning area to pressure
sensors location. Herewith PS1 signal
was ahead of time from PS2 signal
which is associated with location of
sensors. Correlation from different
pressure sensors started to decrease in
about 3 ms after the AD occurrence.
Apparently, it’s associated with
arc movement on the surface of the
electrodes. AD is moving to body
foot towards PS1 under the action of
ponderomotive force.

Fig. 2 shows that the

rst pressure

extremums followed at 0,8 ms
interval, then one-step transition was
down to

≈

0,6 ms. There is some

correlation between the signals from
pressure sensors and the discharge
voltage oscillogram. Thus, the

rst

PS1 pressure maximum is equivalent
of the "smeared" voltage maximum.
There was a voltage step up to
2.2 kV in 3.64 ms after arc ignition
(Fig. 1) before the absolute pressure
maximum in the oil, which was

xed

in 3.71 ms after AD occurrence and
amounted to

≈

1,7 MPa (Fig. 2).

Apparently, there were sound waves
in a liquid under a sharp voltage
decrease (breakdown).

Fig. 4 shows a pressure

oscillogram measured in nitrogen. TO
rises under the in

uence of expanding

gas-vapor bubble, which leads to gas
compression and pressure increase.

Fig. 2. Pressure in TO close to the foot (pressure vs. time)

Fig. 3. Pressure in TO close to “liquid-nitrogen” interface

(pressure vs. time)

Equipment Protection

Page 5

[email protected], www.eepr.ru

According to high-speed shooting the
liquid level rises uniformly up to 0.1 m
of height, than vapor-gas mixture
ruptured to nitrogen area. Typical liquid
rise rate is approximately 10 m/s, which
is signi

cantly less than the speed of

sound in nitrogen, that’s why the process
of compression is adiabatic. Fig. 4 shows
the calculations of Poisson adiabatic
pressure in gas at two values of liquid
rise rate, which determine gas volume
rate. As

gure shows, the calculated and

measured pressure values are consistent
with each other.

The data of liquid rise rate allows to

calculate components of energy balance
in AD. In the experiment

energy

value was 64 kJ and the liquid maximum
kinetic energy was 3.5 kJ. Thus, the
liquid kinetic energy transferred 5—10%
of total energy released in the discharge,
and the main part of the energy was needed for TO heat
and its decomposition.

After discharge in TO the overpressure in nitrogen

“blanket” was at 10—50 KPa level, which is proportional
to gas volume released due to TO decomposition. The
process of gas formation under AD in TO is characterized
by ratio of gas formation

, which is the ratio of

released gas volume to AD energy. According to our data

= 0.11 liters / kJ.

According to our tests approximately the same value

of gas formation under AD (0.1—0.11 liters / kJ) was
calculated for insulating

uids from vegetable oils (rape,

soya, castor). The electric

eld in AD column burning in

vegetable oils is 0.1—0.3 kV/cm.

Carried out experiments allow us to de

ne qualitative

features of AD dynamic effect in TO to HV OFEE body,
the main of which is lack of shock waves in liquid. Perhaps
a shock wave occurs at the moment of initiator explosion,
but it quickly degenerates into a sound wave. The average
pressure rise rate in liquid is 0.3—0.5 MPa/ms. At the
background of increasing pressure of body walls there
are intensive sound waves. According to our data the
maximum pressure of body wall in our experiments was
about 2 MPa. The pressure in the arc burning zone was
slightly higher. The rate of “liquid- nitrogen” interface
due to expansion of vapor-gas bubble was 10—20 m/s.
Estimated that pressure in vapor-gas bubble at such
boundary

uid rate should be 5—10 MPa.

ARCLESS SOURCE OF PULSE PRESSURE

(ASPP)

The research results of AD de

ned the requirements

for arcless source of pulse pressure (ASPP) using for
simulating AD effect in HV OFEE. The pressure pulse in
ASPP is generated under the expansion of jet of powder
gases (JPG) produced due to combustion of explosive

Fig. 4. Overpressure in nitrogen (pressure vs. time)

materials (EM). It is important that duration of pulse
pressure effect should be about 50 ms. This fact excludes
the possibility of EM using for JPG with the necessary
parameters such as hexogen or trotyl. Therefore, in our
experiments we used gunpowder as EM, because it burns
much slower than trotyl. Gunpowder ef

ciency is 3.8

kJ/g, speci

c gas production — 0.9 l/g.

SPG generator was a high-pressure body, where EM

combustion products

ew out from Laval nozzle. The

pressure pulse magnitude and duration can be controlled
by changing nozzle area, EM mass, EM allocation in
combustion body and ignition methods. Experiments
were carried out in the same body as AD experiments.
JPG generator was attached to one of the lower windows,
so that the area of JPG in

uence was the same as under

arching. TO and water were used as working

uids. Heat

rate under EM combustion

was varied in the range of

10—50 kJ.

Pressure was measured at characteristic points of the

body, high-speed shooting of the liquid was recorded
under JPG in

uence. According to measurements jet

pressure at the entrance of the liquid was about 10—
20 MPa for approximately 1 ms. Duration of jet exposure
to the liquid was varied from 20 to 60 ms. A typical
pressure value of body walls was about 1 MPa. Motion
state of "liquid-nitrogen" interface under SPG in

uence

immersed in the liquid was the same as under AD under
the same energy effect. The interface, while remaining

at, was rising at a rate of 10—20 m/s. It should be noted

that there were no big differences between water and
transformer oil response for JPG impact.

The experiments proved the possibility of hydraulic

similarity of liquid

ow under the JPG and AD effect. The

equivalence of the liquid under JPG and AD effect was
achieved by energy equations and duration of exposure.
In this context, JPG generator (ASPP) can be used in

Page 6

August 25–29, France, Paris

explosion protection tests for simulating AD effect in HV
OFEE body.

Fig. 5 presents ASPP calculated for energy of 5 MJ

before the test of "shooting" from ASPP to controlled
container

lled with air. Fig. 6 shows an "oscillogram"

of pressure in the controlled container. This

gure also

shows the calculated pressure values. As it seen, the
experiment is coherent with the calculation. After reaching
the maximum measured pressure value is reduced due
to powder gases cooling. In the model this effect was
not taken into account. This fact explains the difference
between theory and practice over the time.

ASPP can be used for series of tasks aimed to improve

the explosion protection of HV OFEE:
• Test of HV OFEE production samples;
• Examination of explosion protection systems and

devices ef

ciency;

• Elaboration of HV OFEE new constructions with high

level of explosion protection, model testing;

• Basic data acquisition for development and veri

cation

of numerical methods for HV OFEE perspective
constructions and explosion protection systems.

EXPLOSION PREVENTION TESTS FOR

INSTRUMENT TRANSFORMERS

The explosion protection tests using ASPP were carried

out with serial current transformers (CT) and voltage
transformers (EVT). Test results were described in detail
in studies [9, 11, 12]. A brief summary is presented here.
Transformer models, characteristic values of internal short-
circuit currents and the expected AD energy maximum
are shown in Table 1. The calculations of AD energy
maximum were based on fault currents and transformer
design data from manufacturing plants. During the tests,
pressure sensors were installed at different distances from
JPG input inside the transformer. This test was recorded
by high-speed shooting from two mutually perpendicular
directions with a time resolution of less than 3.3 ms.
Area of SPG effect was consistent with most probable
occurrence of AD area. Transformers were

lled with

transformer oil of GC brand during the tests.

The design of tested transformers was considerably

different. In CT manufactured by Russian Federation
the tank with windings are installed in the bottom
where located the porcelain insulator with the pressure
compensator in bellows form (Fig. 4). Bellows provides
compensation of oil volume temperature changes and
protection of internal insulation from humidifying. In
instrument transformers manufactured by Ukraine the
tank with windings are installed in the top. Protective
bellows are located under the tank with windings. CT 330
is equipped with protective membrane instead of bellows.

110 kV CT

According to test results, exposure time at transformer

overpressure of 0.5 MPa was approximately 60 ms. By

means of high-speed shooting was de

ned

that basic deformation of the body has
occurred during the

rst 20—25 ms after

ASPP ignition, when overpressure inside
CT was about 1 MPa. The transformer
body has undergone considerable plastic
deformations, thus its volume has increased
approximately up to 8.5 liters (12%).
Deformation of the body did not caused
TO leakage.

110 kV Voltage transformer (EVT)

Exposure time of ASPP at energy of

1MJ was 60—70 ms. EVT overpressure
was at 0.5 MPa. In 10—15 ms after ASPP
start the

ange of protective membrane

began to move and

nally torn off. The

bellows has started to move approximately
in 15 ms after ASPP start and it was
completely opened approximately in 20 ms.

Fig. 5. ASPP calculated for 5 MJ

Fig. 6. Pressure in controlled container versus time response

characteristics (pressure vs. time)

Equipment Protection

Page 7

[email protected], www.eepr.ru

№

Transformer

model

Producer,

factory

Fault

current

range, k

Energy

, MJ

Exposure

time of

ASPP, ms

110 kV CT

Russian

Federa-

tion

1.5—3

0.4 60

330 kV CT Ukraine

5—10

3 110 kV EVT Ukraine

15—20

Table 1. Instrument transformers tested by ASPP

The internal pressure maximum was about 4 MPa,
pressure growth rate was about 1 MPa/ms. The protective
membrane was torn off and TO stream level was risen up
to 10 meters. Thus radius of TO disorder was less then
radius of standard safety zone. There were no visible
deformation of transformer metalware and baseinsulator
as opposed to bellows and top guard vessel.

330 kV CT

Tests of two CT models has been carried out. Essential

design defect of the high-voltage block of the tested
model have been revealed during

rst test. The jet of

TO with velocity of 20 m/s started to

ow from service

openings located in this part of CT after ASPP start.
Jet range was about 40 meters. This design of CT has
been rejected. Further results of tests using modernized
design of the high-voltage block are considered. Service
openings in new block have been eliminated, and bellows
instead of a protective membrane has been established.
The considerable plastic deformations of body after tests
have been observed. The maximum deformation of the
transformer body was up to 30 mm, a de

ection of top

ange edges was 20 mm and a swelling of the transformer

top cover was 40 mm. Insulation of transformer windings
has been considerably broken down.

According to accepted de

nition these three tested

transformers including modi

ed CT of TFRM type can be

recognized as explosion-proof and

reproof at AD energy

values from Table 1.

SPECIAL ASPECTS OF TRANSFORMER

EXPLOSION

According to analysis of published studies, the most

vulnerable elements of internal faults occurrence in
power transformers are bushings, oil-

lled cable box and

tap changers. AD develops close to the point of a short
circuit origin between transformer body ("the ground")
and construction elements under high potential. The
length of arc column de

ning AD voltage depends on HV

OFEE construction may range from 0.1 to 0.3 meters.
AD continuously and randomly transfers along internal
surfaces of the transformer due to ponderomotive forces
and convective currents. Since the AD characteristic
average rate is about 10 m/s and the time of its
“lifetime” — ~ 50 ms, the surface area of the transformer
under AD effect is ~ 0,1 m

. Therefore, AD transfers inside

the volume of 10-30 liters. This fact gives a reason for
shock wave absence in HV OFEE despite the high power
of AD.

Pressure equalization period inside transformer tank

is estimated as sound wave double transmission time
of maximum distance between opposite walls of the
transformer. Pressure equalization period in 330 kV
instrument transformers which are similar to previously
discussed transformers is ~ 1 ms. This period is less than
arc duration. Pressure equalization period in distribution
transformers is ~ 15 ms. This means that there is a high

pressure differential in large transformers during arching.
The maximum value is achieved in arching zone. These
calculations show that small transformers should be
destroyed uniformly across the surface under the pulse
pressure in

uence. Such damages were recorded during

our tests of instrument transformers. Failure mode of large
transformers is local where the damaged area is less than
10% of total surface area. An example of such damage
is shown in Fig. 7 — Transformer in Western Siberia
Substation.

The maximum overpressure for transformer body

depends on design, location and pulse duration.
According to general requirements for transformer
body deformation should be in elastic zone with static
overpressure of 0.05 MPa. There is tank rupture under
dynamic loading of the transformer at overpressure
above 0.5 MPa for more than 5 ms.

Apparently most probable conditions for transformer

explosion are in the range of 10—30 ms after AD ignition.
At early stage of AD burning for about 10 ms, internal
pressure of the transformer does not reach critical values.

Fig. 7. Transformer in Substation after explosion

Page 8

August 25–29, France, Paris

At later stages of AD in about 30 ms, the probability
of explosion is lower. Firstly, AD electric power is
signi

cantly reduced due to increase of insulating

uid conductivity. As a result there is a decrease of gas

formation rate in AD and changes in pressure growth
rate. Secondly, there is an effect of internal volume of the
transformer tank due to elastic and plastic deformation
of the walls under high pressure in

uence by this point

of time. The additional volume partially compensate for
pressure growth due to TO decomposition.

In the view of foregoing considerations, we can lay

down basic requirements for explosion protection systems
of HV OFEE:
• the response time for internal pressure increase should

not exceed 5 ms;

• the system should limit the pressure maximum in the

transformer tank at the level of 0.3—0.5 MPa.
Protection system should be installed close to problem

areas of the transformer in case it is not possible to protect
the entire internal surface.

MODEL TESTS FOR EXPLOSION

PROTECTION OF HV OFEE

Well-known protection methods of HV OFEE focus

on formation of additional volume

V for expansion of

TO in case of internal pressure growth under AD. The
effectiveness of the protection system can be estimated
from:

V/ B

(1)

The

value so-called a protection system reliability

parameter is the ratio of the additional volume of TO and
gas volume released due to TO decomposition under AD.
Approximately we can take as follows
•

> 0.7—0.8 tank deformation is elastic and the

equipment is explosion safe as well as explosion proof;

• 0.3 <

<0.7 should be considerable plastic deformation

of the transformer body;

•

< 0.1—0.3 should be expected explosive destruction

of the body.
Tentatively we can take typical values of protection

system reliability as

~ 0.7—0.8

~ 0.1—0.3.

There are two ways of additional volume formation

for TO in literature. The

rst method is based on

using porous coverings on the internal surfaces of HV
OFEE body [13]. It is expected that porous material
is compressed under high pressure in

uence resulting

formation of available volume, which is

lled with

expanding hydrogen, carbon oxides and hydrocarbon
gases due to decomposition of TO. In consequence
pressure growth inside the body is limited. Additional
protective effect can be achieved in case compression
of material is enough to spend a signi

cant portion of

kinetic energy of TO

ow for compression. "Porous

covering" method may be effective in case substantial
compression of porous material takes place at a relatively
low overpressure — approx. 0.3—0.5 MPa. Up to the
present day this method of protection has not passed the
tests and it’s not used in practice.

The second method of protection is so-called "protective

membrane" method. Principle of method is installation
of protective membranes on HV OFEE body which are
destroyed under action of AD pulse pressure and used for
TO

ow to special container. [14] It is considered that this

way internal pressure of HV OFEE can be kept within
acceptable limits. The second protection method is widely
used, for example in SERGI Transformer Protector system.

Porous covering method

In the experiments of this protection method we used

HV OFEE breadboard model which looks like steel
cylindrical tank of 0.95 m

volume and 1.45 m height

with a conical nozzle at the bottom. Tank diameter was
1 m, thickness was 7 mm. The cover has been screwed
to a sidewall by 24 bolts with thread diameter of 12 mm.
The plate of polyfoam of 50 mm thickness has been glued
on a steel cover of a breadboard model. The polyfoam
was made from extruded crumb with 0.04 kg/dm

density.

The tank was

lled with water. ASPP was installed at a

distance of 0.2 m from the top cover. Estimated value of
ASPP pulse was 0.35 MJ.

Figure 8 shows a shot with breadboard model in 80 ms

after ASPP start, when the cover has lifted in 0.8 meters.
After experiment only 3 bolts from 24 has survived, the
cover de

ection has amounted to 50 mm, the polyfoam

has crumbled into small fractions.

Hence the discussed protection method of porous elastic

covering cannot protect OFEE body from considerable
deformation under action of a high pressure pulse. This
result was expected. Indeed for the effectiveness of this
protection system needed volume increase which is
accessible for the liquid due to compression of porous
material for its compensation in 3—5 ms — the time of
pressure rise in the liquid. This is possible either at slow
pressure rise rate of 0.1 MPa/ms, or at small sizes of the

Fig. 8. Test of HV OFEE breadboard model with

porous covering

Equipment Protection

Page 9

[email protected], www.eepr.ru

protected model of 0.1 meter. The
main impulse load took only a part
of tank working surface under AD
and internal volume increased due to
compression of porous material was
not enough to prevent an explosion.

These tests were carried out using

polyfoam as a damper material.
It’s possible to use other damper
materials with different properties.
However, the ef

ciency of the

protection system in general cannot
be suf

ciently high. This fact can be

illustrated by following calculations.
If the typical size of transformer tank
is

, then the possible increase of

available oil volume under covering
of transformer internal surface
with damper material of maximum
compression

will be about:

(2)

The porous covering regularly

compresses under action of pulse
pressure only in case tank size is under

~ 0,5 m. Taking for calculations that

~ 0,5 m and

~ 0,02 m, we

nd by

formula (2),

~ 30 liters. It may be

enough to protect the transformer from the explosion at
AD energy of 0.5 MJ (

~ 55 liters).

At the increasing size of transformer tank after

short circuit occurrence the covering can be effectively
compressed only close to the short-circuit origin with total
area of 1 m

. In this case, the additional volume will be

~ 20 liters. The effectiveness of this protection system

under AD energy

~ 1 MJ (

~ 110 liters) will be

~ 0.2, i.e. even at a relatively low AD energy we can

expect explosive destruction of tank.

In summary, porous coverings which are compressed

at pulse pressure of 0.3—0.5 MPa, having an effective
Young's modulus of 0.5 MPa, may be used in explosion
proof transformers with the tank size up to 0.5 meters if
expected AD energy does not exceed 0.5 MJ. In addition
this porous covering material has to maintain its properties
during continuous operation.

Method of a protective membrane

Fig. 9 shows a test of OFEE breadboard model using

this protection method. The HV OFEE model looks like
steel cylindrical vessel of 1,4 m diameter

lled with water.

The thickness of breadboard model steel cover is 12 mm.
The air volume has been separated from the water by an
aluminum membrane with the thickness of 0.2 mm and
diameter of 200 mm. The concrete blocks of 70 kg have
been installed at a distance of 300 mm from the cover to
simulate transformer windings. Area of ASPP effect was
between concrete blocks and cover at 0,2 m distance from
roof opening — opposite the PS2. PS1 detected pressure

in air bubble behind the membrane, PS3 and PS4 detected
pressure in

uid from far

eld of ASPP effect. The pulse

action for HV OFEE body was recorded by high-speed
shooting. Pulse energy of high pressure was 1MJ, pulse
duration was 50 ms.

Membrane contact sensor recorded its rupture in 3 ms

after ASPP start. The water

ow rate through membrane

calculated due to air pressure changes in air bubble
behind the membrane was 20 m/s. Pressure maximum in
the liquid has reached 1.8 MPa. High-speed shooting has
shown that the deformation of HV OFEE body had lasted
for 10—15 ms. After the test we found out that the residual
deformation of steel cover was about 40 mm and concrete
blocks were moved for 50 mm. The conditions for the
discussed protection method in this test were optimal: a
thin membrane of a large diameter was installed right in
front of the epicenter of the pressure rise. However the
present protection method was not effective enough.

The tank rupture prevention system in Fig. 9 is a

simpli

ed version of SERGI Transformer Protector (TP)

system. This protection system is used in energy utilities of
Russia in recent years, but the experience of its operation
is not encouraging. On September 22, 2009 there was an
explosion of AT-1 — 330 kV tank due to ISC at substation
"Mashuk", where this system was installed. SERGI gave
an explanation of this rupture in the report [15]. According
to this report there was a peak current of 10 kA and arc
duration of 60 ms at the time of rupture. In the analysis
of TP system SERGI experts assumed the AD voltage
of 37 kV, so that the energy released in AD was about

Fig. 9. HV OFEE breadboard model for tests using protective

membrane method

Page 10

August 25–29, France, Paris

11 MJ. This AD voltage value seems conservative, because
it was calculated without taking into account the voltage
loss in inductance. According to our estimates, the AD
voltage was signi

cantly low, so that the total AD energy

was about 4 MJ. GPFD volume (gaseous products of

uid

decomposition) was about 0,45

under this

energy.

According to data [15] the protective membrane of

8 inches diameter (~ 250 mm) was destroyed in 4,5 ms
after short-circuit occurrence at pressure of 0.08 MPa.
There was TO

ushing through the opening which

caused "Depressurisation" of transformer tank in 112 ms.
According to this report [12], even though the TP system
did not protect the HV OFEE body from explosion it
prevented the

re occurrence. However, the data in the

report [15] raise some doubts. According to this data the
maximum oil

ow rate through the destroyed membrane

does not exceed 20 m/s. According to calculations based
on report values of TO

ow rates, approximately 25 liters

of TO were leaked through the diaphragm under AD.
Therefore, the reliability coef

cient of the protection

system (1) is

~ 0.07, that’s why TP couldn’t protect HV

OFEE tank from explosion. As for the lack of

re, the

probability of its occurrence after rupture does not exceed
20%, so we cannot give a credit to TP system especially as

energy was relatively small.

For information, the autotransformer tank rupture after

explosion on "Mashuk" substation was local, total damage
area was 1

TESTS OF DYNAMIC PROTECTION

SYSTEM

The analysis shows that well-known explosion

prevention systems of HV OFEE are not effective enough
and it’s a necessary to develop new protection systems.
This section brie

y describes tests results of dynamic

protection system (DPS) developed in Joint Institute
for High Temperatures, Russian Academy of Sciences.
The main elements of DPS are spring-loaded running
blocks (Fig. 10). The blocks were installed on the side of
transformer body close to the most probable occurrence
of short circuit area. The maximum displacement of the
blocks under in

uence of pulse pressure was ~ 0,3 m.

Protection of bushings was carried out
using special diaphragms.

The tests were carried out in

autotransformer (AT). The autotransformer
was out of service, but all the elements
have been preserved inside the body. Fig.
11 shows a photo of the transformer with
established DPS elements (guard vessels are
painted blue). There were 16 running blocks
under circular guard vessel and 35 blocks
under rectangular guard vessel (Fig. 10).
There was an ASPP inside the protective
chamber from the left. DPS wasn’t installed
at the back side of the transformer, it was
installed at one of three bushings.

ASPP with energy of 1 to 3 MJ and

exposure duration of 30 to 50 ms was
used in the tests. High-speed shooting
(up to 2000 frames per second), four PS
and displacement sensors were used for
diagnostics of the tank deformation. The AT
tank was

lled with water.

A series of experiments were carried

out. The pressure pulse was supplied to
most likely points of short circuit origin
from both sides of transformer, including
the bushing area. Plastic deformation of
the tank with partial destruction of the
structural elements but without leaks was
recorded under pulsing of the back side of
autotransformer without DPS. High-speed
shooting has recorded that blocks started to
move in 5 ms after pulsing.

The study highlights:
• pressure maximum in autotransformer

breadboard model increases approximately

Fig. 10. Dynamic protection system (working valve

block without guard vessel)

Fig. 11. AT with DPS before tests

Equipment Protection

Page 11

[email protected], www.eepr.ru

in proportion to ASPP energy: under the energy of 1 MJ
pressure maximum is about 0.5 MPa, under the energy of
3 MJ the pressure exceeds 1 MPa.
• The basic body deformation without DPS starts in

20—30 ms after the pulsing.

• Displacement rate of blocks with DPS increases due

to ASPP energy increase: maximum velocity of blocks
reaches 30 m/s at ASPP energy of 3 MJ.

• DPS has much lower response time in comparison with

the factory explosion protection system of protective
membrane.

• DPS installed in front of the pulse pressure bushings

protects the body from plastic deformation up to 3 MJ
of pulse energy.
It is estimated that DPS with tested con

guration has

reliability coef

cient

~ 0,5. The reliability of explosion

protection can be increased from 30 up to 50% in case
DPS is installed both on sides of the transformer and all
high-voltage bushings.

CONCLUSIONS

1. Presented research results of transformer oil

and liquid

ow under action of AD and jet of powder

gases (JPG). The study shows that the transformer oil

ow and liquid

ow under action of AD and JPG impact

are similar at equal values of energy duration and energy
exposure.

2. Developed an arcless source of pulse pressure

(ASPP) for

re and explosion prevention tests of HV

OFEE under energy of 5 MJ.

3. Developed test method for HV OFEE using ASPP.

The proposed method is recommended for model tests
and for evaluation of explosion prevention system
performance for all types of HV OFEE. This method is
used as an alternative to existing method based on AD
initiation inside the equipment.

4. Described experience of explosion tests for serial

instrument transformers.

5. Experimentally shown that the protection method

of porous coverings which can be compressed under
high pressure of expanding vapor-gas bubble on internal
surfaces of HV OFEE cases — cannot be effective for
large transformers.

6. Test models of “protective membrane” method

showed that this method didn’t protect HV OFEE body
from plastic deformation to its explosion. Reliability
parameter of this protection method doesn’t’t exceed 0.1.

7. Presented dynamic protection system (DPS) of

explosion prevention for HV OFEE. Tests of DPS installed
inside autotransformer showed that DPS prevents the
explosive destruction of the autotransformer body at least
at the energy of 3 MJ.

LITERATURE

1. Darian L.A., Arakelian V.G, Gas-resistance of

insulation liquids. Electrotechnika (rus)

№

2, 1997, p.

45—49.

2. Vanin B.V., Lvov U.N., Neklepaev U.N. etc, About

faults of 110—500 kV transformers in service // Power
stations. 2001.

№

9. p.53.

3. Petersen A.,«The Risk of transformer

res and

strategies which can be applied to reduce the risk» //
Session CIGRE-2010. France. Paris. 2010. Report A2-
101.

4. Foata M., Dastous J.B.,«Power transformer tank

rupture prevention // Session CIGRE-2010. Paris.
2010. Report A2-102.

5. Darian L.A., Dementiev Yu.A., Efremov V.P.,

Polistchook V.P., Shurupov A.V., Alternative method
of estimation of explosive safety of high-voltage oil-

lled electrical equipment, Electro, N 5, P. 43—46,

2009.

6. Daryan L.A., Kozlov A.V., Luzganov S.N., Povareshkin

M.N., Polistchook V.P., Shurupov A.V., Shurupova N.P.
Experimental study of a

ow of liquid under action of

an arc discharge and jet of powder gases// Physics of
Extreme States of Matter- 2010. Chernogolovka. 2010.
P. 112.

7. Daryan L.A., Kozlov A.V., Kotov A.V., Povareshkin

M.N., Polistchook V.P., Shurupov A.V., Shurupova
N.P., Pulse arc discharge in mineral and organic oils//
Proceedings of Int. Conf. on Physics of Extreme States
of Matter-2012. 1—6 March, Russia, Elbrus. Institute
of Problems of Chemical Physics. Chernogolovka.
2012. P. 168.

8. Darian L.A., Kozlov A.V., Luzganov S.N., Povareshkin

M.N., Polishuk V.P., Shurupov A.V. Shurupova N.P.,
Pulsed arc discharge in mineral and organic oils //
Proceedings of Low Temperature Plasma Physics —
2011. Petrozavodsk. 2011. Publisher Petrozavodsk
State Universitety. Vol. 1, p. 106—111

9. Darian L.A., Kozlov A.V., Povareshkin M.N., Polishuk

V.P., Shurupov A.V., Son E.E., Fortov V.E., Arcless Tests
of the High Voltage Oil-

lled Electrical Equipment on

Explosion-proof // News of RAS. Energetics. 2011.

№

p. 74.

10. Physical quantities. Catalogue edited by Grigorieva I.S.

and Melikhova M.: Energoatomizdat. 1991. 1260 p.

11. Darian L.A., Polishuk V.P., Shurupov A.V., Arcless Tests

of the High Voltage Oil-

lled Electrical Equipment on

Explosion-proof. Energo-info (rus), 2012,

№

9, p.54—

58.

12. Darian L.A., Kozlov A.V., Povareshkin M.N., Polishuk

V.P., Shurupov A.V., Arcless Tests of the High Voltage
Oil-

lled Electrical Equipment on Explosion-proof //

Electro. 2011.

№

5. p. 23.

13. Mishuev A.V., Kazennov V.V., Gromov N.V., Fire and

explosion protection device for transformer under AD.
Patent RU 2334332.

14. Manie F., Explosion protection device for transformers.

Patent RU 2263989.

15.

SERGI company. Substation «Mashuk». Activation
of Transformer Protector on 22/09/ 2009. Reference
Are61rr00.03e

от

09/10/09.

Оригинал статьи: Tеst Models for Explosion Protection of High Voltage Oil Filled Electrical Equipment

Читать онлайн

Life time of transformers or another HV oil-fi lled electrical equipment (HV OFEE) is about several decades. The gradual degradation of paper-oil insulation occurs under the infl uence of partial discharge, heating, cavitation and other factors in service.