Some problems in an assessment of the consequences of a fire and an explosion during the multicompon

1
Some problems in an assessment of the
consequences of a fire and an explosion during
the multicomponent mixture unknown composition
release.   Melania
Pofit-Szczepanska The Main School of Fire
Service, Firefighting and Rescue Operation
Department Slowackiego Street 52/5401-629
Warsaw, Poland   
2
Consequences of hazardous substances releases
The important part of the safety report is the
analytical part in which fire, explosion and
toxicity hazards are analyzed as well as the
consequences of these releases to the atmosphere.
If the released medium is the substance of a
known composition and the parameters of process
are known too, the calculation of the
consequences of these releases is not difficult.
The descriptions of the way of procedure can be
found in the literature. If however, the mixture
of unknown composition leaks through the rupture
of the pipeline or vessel the assessment of the
consequences of these releases is more
complicated.
3
The method of thermodynamic substitute

• The method has been applied to the assessment of
  fire and explosion consequences during the
  rupture of the pipeline and the release of the
  slops mixture of unknown composition to the
  atmosphere in a Polish refinery.This method
  called the method of thermodynamic substitute,
  is one of the methods used in the calculations
  different parameters dealing with release of
  dangerous substances.

4
The method of thermodynamic substitute

• The single component as a thermodynamic
  substitute is used very often. Of course, the
  most reliable use of a single component
  consequence model results when the single
  component simulates the behaviour of the
  multicomponent fluid over all potential
  conditions from storage conditions to ambient
  atmospheric conditions. Naturally, this involves
  an intimate knowledge of the thermodynamic
  behaviour of the mixture.

5
Single component model
The single component models are used very often
because they are much easier and generally run
faster. Considering D.W. Johnsons example the
release of methane pentane vapour from a large
vessel operating at 3 bars and 65oC and a release
of pure propane in the same conditions. The
release will escape through a 15 cm diameter hole
made at the side of the vessel. The release is
angled 45o above horizontal. The release rate
will be relatively constant since pressure and
temperature in the large vessel will change
slowly with time. In table 1 the computed results
are given and fig. 1 shows the LFL contours. With
many uncertainties the agreement is good. Thus it
appears possible to model release.
6
Fig. 1 Dispersion of vapour releases to the lower
flammable limit
Table 1.
7
Fig. 2 The arrangement of the installation and
the location of vessels
8
Fig. 3 Visible damage on the pipeline which cause
the realistic accidental spill
9
Case study

•  
• Problem
• The refinery, division-slops installation. An
  arrangement of the installation and the
  cylindrical vessels 1, 2, 3 of a division are
  shown in the fig 2. The road A, road 1and
  cross-road A1 can be seen. Tank cars waiting for
  the loading of asphalt and their position when
  fire and explosion took place are marked in
  yellow. The area of an enveloped accident
  amounted to about 5000m2. The fill of vessels 3
  18, 2 26, 1 80.
• An assumed course of the incident
• At first it was a small leak through a small
  hole. It was a long time before the slops were
  released through 0.1m diameter hole in the
  following conditions t 50oC, p 0.5 bar. The
  part of slops according to their density were
  absorbed by the concrete bottom of pipelines duct
  or by the thermo-isolation of pipelines. The
  lightest fraction mixed with air generated the
  cloud of slops. Can be assumed that the released
  quantity of slops was larger than 150 tons
  (calculations). Data published in a literature
  indicate that the generation of cloud during the
  realistic accidental spill is possible if the
  flow out is above 100 kg of relative non-reactive
  fuel (hydrocarbons).

10
Case study

• The following conditions were in the time of
  accidental spill
• night,
• F - Pasquille class,
• T - 2,7oC
• Vv 2 m/s.
• In these conditions, the direction of wind had
  less influence on the
• cloud propagation than the buoyant forces. The
  flammable cloud have
• encountered an ignition source probably somewhere
  in the vicinity
• of the tank car 1 /fig. 2/ in the form of low
  energy source /damaged
• electric installation of tank car 1/.In order to
  analyse the development of
• fire and explosion, two variants of the procedure
  have been discussed.
• In the basis on the information received from the
  rafinery was assumed
• that the released mixture had the C1 to C8
  composition and components
• were in equivalent concentrations.

11
Case study
n-butane could be the substitute C1 to C5 of
the fraction  
n-heksane could be the substitute C6 to C8 of
the fraction /liquid fraction/
For C6 to C8
For C1 to C5
from the following causes  

• the relative density of n-heksane 3.0
• the relative density of released
• liquid mixture
  3.4
• Lower flammability limit
• of n-heksane
  1.52
• Lower flammability limit
• for the mixrure
  1.61

• the relative density of n-butane 2.0
• the relative density of released
• gases mixture
  1.68
• Lower flammability limit
• of n-butane
  2.21
• Lower flammability limit
• for the mixrure
  2.36

12
Case study

• Assumptions applied in calculations
• - fuel-air mixture burns in the way that no
  damaging overpressure is generating /flash fire/
• - generated explosion is the deflagration
• - dispersion of the released mixture occurs in
  two types of surroundings
• a) in obstructed environment, the vapour cloud is
  located in space between dike area and vessels
  and between the vessels 3 and 2 and pipelines
  ducts /fig. 2/, where the ruptured pipeline /fig.
  3/ was situated
• b) in open space the flammable mixture covers
  the area about 5000m2 the area of pipelines duct
  cross-road A1 a space outside of the
  cross-road the space round tank cars waiting
  for the loading of asphalt. 

13
Case study

• The generated cloud was spreading down from SE
  direction to NW.
• - The area of the tank car was ?162m2 (18m x 3m x
  3.5m).
• - The volume 567m3.
• - The area of pipelines duct was about 460m2 (46m
  distance from the realistic accidental leak to
  cross-road marked A1/fig 2/.
• - The open area (non-built) of accident was
  1500m2.
• In tables 2-7 can be seen some results of the
  calculations an overpressure, positive-phase
  duration of explosion and the energies of
  combustion at different

14
Case study

• Distances from accidental leak 10m, 30m, 48m,
  68m and 100m were considered and they made
  calculations of the explosion parameters possible
  in the following places-
• - 10m the nearest distance from an accidental
  leak to the dike area
• - 30m the distance from an accidental leak to
  the vessel 3
• - 48m distance from the accidental leak to the
  cross-road A1
• - 68m distance from the accidental leak to the
  place where probably the piloted ignition
  occurred
• 100m the distance from the accidental leak to
  the place faraway ?30m outside of the cross-road
  A1 /near tank car 1/ /fig. 1/
• On the basis of the technical documentation the
  mass of the accidental leaks was determined
• - 6,000 kg
• - 10,000 kg
• - 20,000 kg

15
EXPLOSION PARAMETERS OF RELEASED SLOPS IN A
FUNCTION OF THE DISTANCE FROM AN ACCIDENTAL LEAK
Table 2. N-butane thermodynamic substitute,
obstructed environment.
 
16
Table 3. N-hexane thermodynamic substitute,
obstructed environment.
17
Table 4. N-butane thermodynamic substitute,
open, non-built environment.
18
Table 5. N-hexane thermodynamic substitute,
open, non-built environment.
19
Table 6. Summary comparison some results of the
calculations of the range cloud vapour explosion
/ST n-hexane/
Table 7. Summary comparison some results of the
calculations of the range cloud vapour explosion
/ST n-butane/
20
The analysis of results
In the basis of the technological data and the
analysis of the run of fire which had taken place
before the explosion, you can explain the
relationship fireexplosionconsequences. The
fire of slops cloud have started at about two
oclock at night when the cloud had already
propagated for about 70 m towards the three tank
cars. About 2 min after fire the first explosion
took place which almost completely damaged the
vessel no.3 filled only in 18 /photo 1-2/.
21
Photo 1. Deformation of vessel no.3 after the
explosion with the visible displacement
Photo 2. Deformation of vessel no.3, visible
detachment from the foundations as a result of
an explosion
22
The analysis of results

• The vessel no.2 /filled 26 / was damaged by the
  second explosion after 45s later /photo 3/.
• Probably, this time was needed to form explosive
  mixture because the hydrocarbons have the narrow
  flammability limits /about 110/. The vessel
  no.3 filled with benzol recovery oil in 80 was
  less damaged /photo 4/.

23
Photo 3. Deformation of vessel no.2, visible
detachment from the foundations and displacement
outside concrete wall
Photo 4. Deformation of vessel no.1
24
Conclusions

• The safety reports and firefighting-rescue plans
  should contain
• characteristics of hazardous materials
• indentification of the threat sources
• the probable scenarios of the accidents
• the quantitative evaluation of the potential
  results for both people and the environment
• To calculate the results of the hazardous
  occurrences the knowledge of the input data is
  needed
• Simulation via the use of pure component
  consequences makes it possible to predict more or
  less accurately conditions when a liquid or gases
  is released.

25
Thank you very much for your attention