Professional Documents
Culture Documents
= ...........................
Equation 3
Where
"
m is the asymptotic burning rate for large pools and k is an effective absorption
coefficient including the mean beam length correction factor. The mass burning rate -
"
m and
the asymptotic burning rate -
"
, ) ( 7 , 0
1
= m k , ) / ( 4 , 46 g kJ H
c
= and 84 , 0 =
chem
x .
Thomas has developed a correlation for the visible length of flame take into account the wind
velocity, which is expressed in equation 4. Where H is the pool flame height (m),
a
ambient
air density (kg/m
3
), g gravity acceleration (m/s
2
) and
*
u nondimentional wind velocity given
by equation 5.
w
u is the wind velocity (m/s).
14
21 , 0 *
67 , 0
"
.
.
55
|
|
\
|
= u
gD
m
D
H
a
...........................
Equation 4
3 / 1
"
*
. .
|
|
\
|
=
v
w
D m g
u
u
...........................
Equation 5
Thomas gave the following correlation for flame tilt based on data from two dimensional
wood cribs.
49 , 0
". .
7 , 0 cos
(
(
(
(
(
|
|
\
|
=
a
w
D m g
u
...........................
Equation 6
Wind can significantly increase the effective pool diameter and corresponding mass burning
rates. The wind tends to both tilt and increase the flame diameter in the downwind direction.
Mudan and Croce suggest the following correlation to estimate the increase in flame diameter.
But, equation 7 should be used with caution at very large wind velocities or small pool
diameter due to a possible flame blow off.
48 . 0 069 . 0
2
25 . 1
|
|
\
|
|
|
\
|
=
a
v w w
D g
u
D
D
...........................
Equation 7
Where
w
D is the effective flame diameter in the presence of wind,
w
u is the wind velocity
(m/s) measured at an elevation of 10m,
v
and
a
are the densities of vapour and air,
respectively.
The thermal radiation flux from a pool fire can be estimate by a point source model or a solid
flame model. The point source model removes most of the geometrical parameters from the
calculation. It assumes that all of the radiative energy from the fire is emitted at a single point
rather than distributed over an idealized shape meant to represent the fire. It requires an
estimate of the total heat release rate-HRR of the fire, and the fraction of that energy that is
15
emitted as thermal radiation. If atmospheric absorption effect are negligible, and the target
distance is large compared to the flame height the radiant point source approximation
provides an attractive simplification to obtain the incident heat flux on the target, equation 8.
2
"
. . 4 x
Q
q
r
= ...........................
Equation 8
In a solid flame radiation model the thermal radiation flux,
"
r
q ,from a fire to a nearby object
is given by equation 9. Where F is a geometric view factor that intercepted by the receiving
object, i.e. target. is the atmospheric transmissivity to thermal radiation, it is a function of
humidity and the distance between the radiation source and receiver.
f
is the effective
emissivity of the flame, expressed by ( )
D
f
e
=
1 where and D are the attenuation
coefficient and pool diameter respectively.
f
E is the total emissive power of the flame at the
flame surface.
f f r
E F q =
"
...........................
Equation 9
For pool fires greater than a few meter in diameter the effective emissivity-
f
is
approximately equal to one. If the atmospheric transmissivity- was taken as one equation 9
become
f r
E F q =
"
, which is the equation proposed by STD 979-20NN from the IEEE.
The thermal impacts caused as a consequence of a transformer pool fire with about 40,000
litter of mineral oil were estimated using the equations mentioned in the previous paragraphs
(i.e. equation 1 to 8). The radiative energy in target from the pool fire of various pool
diameters is presented in Table 4. It was assume that the point source model will provide
reasonable results for target double the flame height (i.e. 2H meter) away from the fire. A
pool fire of 5 meter in diameter is superimposed over a power plant layout, Figure 5.
The high radiate energy from a transformer pool fire increase the temperature of the structures
nearby, therefore its strength and stiffness could be reduced. This effect may lead to
unacceptable deformations or structure failure, such as take out of service a substation
because either the high voltage bar, Figure 4, or the rely house, Figure 6, suffer irreversible
16
thermal damages. At about 70
o
C damage to electronic components are irreversible. What
could happen if a transformer fire causes irreversible damage to the electronics equipment
placed inside the rely house (Figure 6), either due to the transformers close proximity to the
house, or difficulties in controlling and extinguishing it before the heat transfer from flame
takes the substation out of the network? In this case, about 2 million people would be without
electricity for quite some time. Imagine hospitals without electricity, a big city without traffic
lights, and so on.
In a relay house some of the protection, bypass and control circuits which control the
transformers and other substation equipments are placed. In the substation shown in Figure 6
there are two relay houses. Their structure is masonry with glass windows. The structural
thermal performance simulation of one of the relay houses during a transformer pool fire was
carried out by use of the finite elements method. The distance between the transformer and
the relay house is 3.50 meters. Ninety minutes after the fire was initiated, the wall temperature
distribution showed some wall temperatures to be above 300
o
C. Such high temperatures can
cause wall failure. Despite this, the structural integrity of the walls, pillars and beams occurs
only 117 minute after the fire is started. On the other hand, 45 minutes after the fire is
initiated, the glass windows will break. Therefore there will be a rapid growth of the gas
temperature inside the relay house. The heat impact will cause irreversible damage to the
electronic panels. The gas temperature distribution analysis inside the house showed that
temperatures higher than 70
o
C could be reached in approximately 30 minutes.
17
TABLE 4. Radiate energy (kW/m) in a target from a transformer pool fire.
Pool Diameter and Radiate Energy-kW/m
2
Distance to the Target
m 1 m 2 m 3 m 4 m 5 m
1 -- -- -- -- --
2 3,6 -- -- -- --
3 1,6 -- -- -- --
4 0,9 5,4 -- -- --
5 0,6 3,4 -- -- --
6 0,4 2,4 6,3 -- --
7 0,3 1,7 4,6 -- --
8 0,2 1,7 3,5 6,7 --
9 0,2 1,1 2,8 5,3 --
10 0,1 0,9 2,3 4,3 7,0
11 0,1 0,7 1,9 3,5 5,7
12 0,1 0,6 1,6 3,0 4,8
13 0,1 0,5 1,3 2,5 4,0
14 0,1 0,4 1,1 2,2 3,5
15 0,1 0,4 1,0 1,9 3,0
18
Legend
Zone Energy Target Distance
> 5 kW/m 11 metros
> 2kW/m 18 metros
FIGURE 5. A pool fire transformer superimposed on step by transformer bay.
19
FIGURE 6. Detail of the layout of the substation.
A number of workers have correlated threshold of pain and blistering due to a thermal impact,
Table 5 shows some limits to pain and injury given in the literature. The energy liberate from
the transformer fire illustrated in Figure 5 is on the threshold of pain.
TABLE 5. Threshold to pain.
Thermal Radiation Intensity
1,5 kW/m Threshold of pain
2,5 kW/m Level at which pain is felt after 1minute
1 kW/m Level just tolerable to a clothed man
8 kW/m Level which causes death within minutes
4,7 kW/m Threshold of pain. Average time to
experience pain 14.5 seconds
When the load bearing capacity becomes equal to the applied load there will be the structure
failure. During the present study was carried out two structural analysis of an unprotected
steel member (i.e. tension and compression member) as presented below. The analysis was
based on the EUROCODE 3 part 1-2. It suggests a simplified method to estimate the
temperature and the load bearing capacity. In other words, the temperature is assumed
20
uniform over the cross section, and it does not take into account the effects of the restrictions
of the real structures.
Case 1- Tension member:
o Unprotected steel section MR250 steel;
o I cross section 152x18.6;
o Length of member: 4,5 m
o Tension effort in member: 224 kN;
Case 2- Compression member:
o Unprotected steel section MR250 steel;
o 2 C cross section 305x30.8;
o Length of member: 4,0 m
o Tension effort in member: 200 kN;
Possible transformer pool fire radiate energies as a function of the time to the structure
member fails and its temperature for both case 1 and 2 are relate in Table 6.The Figures 7, 8
and 9 show the temperature versus time and the load bearing capacity versus time graphs to
19kW/m
2
, 20kW/m
2
and 7kW/m
2
, respectively.
TABLE 6. Time to fail of the structure member of cases 1 and 2.
Case 1: Tension Member
Section Factor: 271 m
-1
Case 2: Compression Member
Section Factor:175 m
-1
Energy
(kW/m) Temperature
(C)
Failure Time
(min)
Temperature
(C)
Failure Time
(min)
20 10 11
18 12 13
16 16 15
14 NF 18
12 NF 23
10 NF 35
08
507
NF
415
NF
21
FIGURE 7. Case 1: Temperature versus time and load bearing capacity versus time graphs to
an exposed energy of 19kW/m.
FIGURE 8. Case 2: Temperature versus time and load bearing capacity versus time graphs to
an exposed energy of 20kW/m.
22
FIGURE 9. Case 2: Temperature versus time and load bearing capacity versus time graphs to
an exposed energy of 7kW/m.
The structure failure depends on fire severity, steel area exposed to the flames, the applied
load and support conditions. The tension member failures (case 1) are associated with
energies higher than 16kW/m
2
. The compression member failure is probable with energies
around 10kW/m
2
. For the fire simulates, Figure 5, either tension or compression members
failures are not foreseen, but scenarios are subject to uncertainty.
The basic of fire hazard assessment is a set of scenarios. In this paper only the scenario of
pool fire transformer was analyzed. Both completeness and realism of these scenarios are
related to the release itself and to the escalation both are subject to uncertainty, Figure 10.
There is a further uncertainty concerning to the geometry of the release. The transformer oil
will not necessarily come out as a pool, but may issue as a spray fire and may impinge on
other equipment.
An autotransformer of 150MVA protected by a spray system caught fire as a result of a bushing
failure, Figure 10. The spray system did not operate when the fire began. This delay cause the fire
spread, due to the oil leakage through the base of the bushing. NFPA 15 describes water spray design
for transformer. Even if the suppression systems are available there are many uncertainties for their
success or failure in controlling the fire. Questions such as, Can water discharge from the spray
system? Can water terminate the fire? Water will be discharged from the spray system if all water-
supply valves are open when the sensor fuses will enough water reaches the spray head? On the other
hand, the violent nature of a transformer fire could render automatic spray system useless. Although
23
this may have occurred in may cases the automatic spray system did survive the explosion and were
credited with controlling fire, limiting damage and minimizing system (i.e. a substation or a power
plant) downtime. As the system in which the transformer is insert has a dynamic behavior, the fire
scenarios are subject to many uncertainties.
Figure 10. Escalation of a transformer fire.
FINAL CONSIDERATIONS
Power plants and substations have been around for quite some time, so ample engineering
experience exists and the public is familiar with their equipment and structures (i.e.
transformer, circuit break, transmission lines, etc). They also have a substantial economic
incentive to prevent accidents. In spite of mature technology, good management, and
incentives to keep the plant or substation from blowing up, uncontrollable fire rages within
them on occasion, killing operators and causing substantial losses.
Fire in substations range from those which have a relatively minor impact, in which there is
little or no interruption of the operation to the interconnect network to major catastrophe: the
blackout in Southeast of Brazil in 1995 being synonymous. While the engineers who design
the substation have the knowledge and understanding to recognise the fire hazard throughout
the system interactions and take measures, which will reduce the risk of a fire occurring, it is
the substation operators who are responsible for its safe operation on a day-to-day basis. They
Time t
2
t
2
24
must be aware, not only of the inherent hazard of the process of which they are in charge, but
also of what can go wrong and, perhaps more importantly, how it can go wrong.
An insurance company, a fire officer or an industrial company organization all have different
ways of dealing with the fire risk. It involves understand what is at risk, have a sense of the
fire severity and how to dealing with it. In other words, we should answer questions such as:
What is the risk? How serious is it? And what are the alternatives for dealing with it? Some
view of fire risk management in Brazil focus only on a decision of what type of insurance to
purchase. Some organizations ignore the risk in the hope that misfortune will not happen to
them. Based on the assumption that if a transformer fire occurs decisions of what to do will be
handled at the time. An engineering of a suitable protection measures for a suitable protection
for a transformer fire should involved three steps: scenario identification, consequence
analysis and protection evaluation. Some considerations about transformer fire scenarios and
its consequences were discussed in this paper.
25
BIBLIOGRAPHY
Dietrich, W et alli CIGRE working group 5. An international survey on failures in large
power transformer in service. Electra, 88:21-42. 1983.
Duarte et al. Technical Report on fire in substations prepared by the Laboratory of
Technology Risks of Federal University of Pernambuco. RISCTEC-UFPE, Centro de
Tecnologia. Cidade Universitria, Pernambuco, Brazil. 1999.
Granttan, K., Baum, H., Hamins, A. Thermal radiation from large pool fire. NISTIR
6546. November 2002.
LEES, F. P., Loss Prevention in the Process Industries. 2Rev.ed. Oxford, Butterworth-
Heinemann, 2001.
EN 1991-1-2. EUROCODE 1. Actions on structures part 1.2: general actions actions
on structures exposed to fire. CEN, 2002.
Medina, M., Mohallen, C. A. and Duarte, D. (2003). Failure of power transformer: A
contribution on its analysis and reduction. In the proceedings of the Tenth Meeting of
CIGR for Latine America X ERLAC. Argentina, May 18-22, 2003.
ZALOSH, R. G., Industrial Fire Protection Engineering, Worcester - EUA, John Wiley
&Sons, 2003.
Duarte, D. A performance overview about fire risk management in the Brazilian
hydroelectric generating plants and transmission network. Journal of Loss Prevention in
the Process Industries 17, 65-75. 2004.
EN 1993-1-2. EUROCODE 3. Design of steel structures part 1.2: general rules
structural fire design. CEN 2005.
Bastos, G.
Duarte et al . Identification of the transformer failures with a new logic diagram.
Proceedings of the CIGRE Colloquium in Bruges (Belgium) on October 7-12, 2007.
IEEE Std 979-20NN (2007). IEEE guide for substation fire protection. The Institute of
Electrical and Electronics Engineers. Unpublish.
Denis Allan. Fire avoidance in transformer substation. CIGRE UK.