12B.3
Site Description
12B.3.1
Population Data
The BPPS area is generally remote with very low
population in the vicinity. Both
land and marine populations are considered in the analysis, for the year 2011
(the expected year of completion of the 1st GRS), the phase 2 construction period in 2014 and a future year 2021. The consequence analysis (Section 12B.6) demonstrates that the
maximum extent of hazard effects reach about 335 m. The population assessment therefore
considered all offsite population within about 500 m of the proposed GRS. The population at the GRS sites, such as
workers, are considered onsite population and therefore outside the scope of
this QRA.
Land Population
There is no land based population within 500 m of the
proposed GRSs.
The security entrance to BPPS is more than 600 m from
the GRS facilities. The nearest
industrial facilities in Lung Kwu Sheung Tan are about 1.4 km away and Lung Kwu Tan Road
is more than 700 m away. None of
these populations will be impacted by any release from the GRSs.
Marine Population Estimation
Black Point is situated near Deep Bay. The marine traffic in the vicinity
includes passenger ferries, container ships and rivertrade vessels going to Guangzhou and other Pearl River Ports. Small fishing vessels and leisure crafts
also contribute to the marine traffic in the Black Point region.
Vessel Population
The vessel population used in this study are as given
in Table 12B.1. The figures are based on BMT’s Marine
Impact Assessment report [3] except those for fast ferries. The maximum population of fast ferries
is assumed to be 450, based on the maximum capacity of the largest ferry
operating in the area. However, the
average load factor for fast ferries to Pearl River
ports is only 31.8% [4]. Hence, a
distribution in ferry population was assumed as indicated in Table 12B.1. This distribution gives an overall load
factor of about 52% which is conservative and covers any future increase in
vessel population. There is an
additional category in the traffic volume data called ‘Others’. These are assumed to be small vessels
with a population of 5.
Table 12B.1 Vessel
Population
|
Type
of Vessel
|
Average
Population per Vessel
|
%
of Trips
|
|
Ocean-Going Vessel
Rivertrade Coastal vessel
Fast Ferries
Tug and Tow
Others
|
21
5
450 (largest ferries with
max population)
350 (typical ferry with
max population)
280 (typical ferry at 80%
capacity)
175 (typical ferry at 50%
capacity)
105 (typical ferry at 30%
capacity)
35 (typical ferry at 10%
capacity)
5
5
|
3.75
3.75
22.5
52.5
12.5
5.0
|
Marine Vessel Protection Factors
The population on marine vessels is assumed to be
provided with some protection from the vessel structure. The degree of protection offered depends
on factors such as:
·
Size
of vessel;
·
Construction
material and likelihood of secondary fires;
·
Speed
of vessel and hence its exposure time to the flammable cloud;
·
The
proportion of passengers likely to be on deck or in the interior of the vessel;
and
·
The
ability of gas to penetrate into the interior of the vessel and achieve a
flammable mixture.
Small vessels such as fishing boats will provide
little protection but larger vessels such as ocean-going vessels will provide
greater protection. Fast ferries
are air conditioned and have a limited rate of air exchange with the
outside. Based on these
considerations, the fatality probabilities assumed for each type of vessel are
as given in Table 12B.2.
Table 12B.2 Population
at Risk
|
Marine
Vessel Type
|
Population
|
Fatality
Probability
|
Population
at Risk
|
|
Ocean-Going Vessel
Rivertrade Coastal Vessel
Fast Ferries
Tug and Tow
Others
|
21
5
450
350
280
175
105
35
5
5
|
0.1
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.9
0.9
|
2
2
135
105
84
53
32
11
5
5
|
Methodology
In this study, the marine traffic population in the vicinity
of Black Point has been considered as both point receptors and average density
values. The population of all
vessels are treated as an area average density except for fast ferries which
are treated as point receptors.
The marine area around Black Point was
divided into 12.67 km2 grid cells, each grid being approximately 3.6
km ´ 3.6km.
The transit time for a vessel to traverse a grid is calculated based on
the travel distance divided by the vessel’s average speed. The average speed [2] and transit time for different vessel types are
presented in Table 12B.3.
Table 12B.3 Average
Speed and Transit Time of Different Vessel Type [2]
|
Type of Vessel
|
Assumed Speed (m/s)
|
Transit
Time (min)
|
|
Ocean-going
vessel
|
6.0
|
9.9
|
|
Rivertrade
Coastal vessel
|
6.0
|
9.9
|
|
Fast Ferries
|
15.0
|
4.0
|
|
Tug and Tow
|
2.5
|
23.7
|
|
Others
|
6.0
|
9.9
|
The number of vessels traversing each grid daily was
provided by a previous marine study [2]. These
are given in Table 12B.4, where the
grid cell reference numbers are defined according to Figure 12B.4. The marine
study was based on 2003 data, extrapolated to years 2011 and 2021.
The number of marine vessels present within each grid
cell at any instant in time is then calculated from:
Number of vessels = No. of vessels per day × grid length / 86400
/ Speed (1)
This was calculated for each type of vessel, for each
grid and for years 2011 and 2021.
The values obtained represent the number of vessels present within a
grid cell at any instant in time.
Values of less than one are interpreted as the probability of a vessel
being present.
Figure 12B.4 Grid
Cell Numbering Scheme
Table 12B.4 Number
of Marine Vessels per Day
|
Grid
No.
|
Average Number of Vessels Per Day
|
|
|
2011
|
2021
|
|
|
OG
|
RT
|
TT
|
FF(*)
|
OTH
|
OG
|
RT
|
TT
|
FF(*)
|
OTH
|
|
1
2
3
4
|
19
0
19
0
|
788
0
557
368
|
368
21
263
168
|
44
0
77
11
|
567
84
294
294
|
23
0
23
0
|
863
0
610
403
|
503
23
288
184
|
52
0
91
13
|
621
92
322
322
|
OG = Ocean-going vessels
RT = Rivertrade coastal vessels
TT = Tug & tow vessels
FF = Fast ferries
OTH = others
(*) Fast
ferries are treated separately
Average Density Approach
The average marine population for each grid is
calculated by combining the number of vessels in each grid (from Equation 1) with the population at risk
for each vessel (Table 12B.2). The results are shown in Figures 12B.5 and 12B.6. This grid
population is assumed to apply to all time periods. As can be seen, the growth in marine
population from 2011 to 2021 is only marginal. Intermediate values are applied to the
phase 2 construction in 2014. Note
however that fast ferries are excluded since ferries are treated separately in
the analysis (see below).
When simulating a possible release scenario, the
impact area is calculated from dispersion modelling. In general, only a fraction of a grid
area is affected and hence the number of fatalities within the grid is
calculated from:
Number of fatalities = grid
population × impact area / grid area (2)
Figure 12B.5 Marine
Population at Risk by Grid, Year 2011
Figure 12B.6 Marine
Population at Risk by Grid, Year 2021
Point Receptor Approach
The average density approach, described above,
effectively dilutes the population over the area of the grid. Given that ferries have a much higher
population than other classes of vessel, combined with a relatively low
presence factor due to their higher speed, the average density approach would
not adequately highlight the impact of fast ferries on the FN curves. Fast ferries are therefore treated a
little differently in the analysis.
In reality, if a fast ferry is affected by an
accident scenario, the whole ferry will likely be affected. The likelihood that the ferry is
affected, however, depends on the size of the hazard area and the density of
ferry vessels. To model this, the
population is treated as a concentrated point receptor i.e. the entire population
of the ferry is assumed to remain focused at the ferry location. The ferry density is calculated the same
way as described above (Equation 1),
giving the number of ferries per grid at any instant in time, or equivalently a
“presence factor”. A hazard
scenario, however, will not affect a whole grid, but some fraction determined
by the area ratio of the hazard footprint area and the grid area. The presence factor, corrected by this
area ratio is then used to modify the frequency of the hazard scenario:
Prob. that ferry is affected = presence
factor × impact area / grid area (3)
The fast ferry population distribution adopted was
described in Table 12B.1. Information from the main ferry
operators suggests that 25% of ferry trips take place at night time (between
7pm and 7am), while 75% occur during daytime. Day and night ferries are therefore
assessed separately in the analysis.
The distribution assumed is given in Table
12B.5.
Table 12B.5 Fast
Ferry Population Distribution for Day and Night Time Periods
|
Population
|
Population
at Risk
|
% of Day
Trips
|
% of
Night Trips
|
% of All
Trips
(= 0.75
× day + 0.25 × night)
|
|
450
350
280
175
105
35
|
135
105
84
53
32
11
|
5
5
30
60
-
-
|
-
-
-
30
50
20
|
3.75
3.75
22.5
52.5
12.5
5.0
|
The ferry presence factor (Equation 1) and probability that a ferry is affected by a release
scenario (Equation 2) are calculated
for each ferry occupancy category and each time period.
Stationary Marine Population
Other stationary marine population such as that for
the Urmston Road Anchorage area are more than 500m from the
proposed GRSs and were therefore neglected in the analysis.
12B.3.2
Meteorological Data
Data on local meteorological conditions such as wind speed,
wind direction, atmospheric stability class, ambient temperature and humidity
was obtained from the Hong Kong Observatory [20], [21].
The location of weather stations in the vicinity of
the GRS is shown in Figure 12B.7. Data from the Sha Chau weather station
was adopted for the GRS study as it is closest to the site and also the most
relevant based on the topography.
The meteorological data used in this study is based on the data recorded
by the stations over a five year period from 2004 to 2008.
Figure 12B.7 Weather
Stations in Vicinity of Black Point
The raw data from the Observatory is a series of readings
taken hourly over the 5-year period.
This data was rationalized into four combinations of wind speed and
atmospheric stability class, denoted as 2.5B, 3D, 7D and 2F where 2F for
example refers to a wind speed of 2m/s and atmospheric stability class F. The data is then further sorted in 12
wind directions. This sorting of
meteorological data is performed for the two time periods, day and night.
The fraction of occurrence for each combination of
wind direction, speed and atmospheric stability for each time period is
presented in Table 12B.6.
Wind directions, such as 90°, refer to the direction of the prevailing
wind. For example, 90° refer to an easterly wind, 0° is northerly, 180° is southerly and
270° is westerly.
Table 12B.6 Data
from Sha Chau Weather Station (2004-2008)
|
|
Day
|
Night
|
|
Wind Speed
(m/s)
|
2.5
|
3
|
7
|
2
|
2.5
|
3
|
7
|
2
|
|
Atmospheric
Stability
|
B
|
D
|
D
|
F
|
B
|
D
|
D
|
F
|
|
Wind Direction
|
Fraction of Occurrence
|
|
0°
|
0.052
|
0.006
|
0.140
|
0.005
|
0.000
|
0.005
|
0.114
|
0.010
|
|
30°
|
0.009
|
0.005
|
0.050
|
0.003
|
0.000
|
0.004
|
0.092
|
0.008
|
|
60°
|
0.005
|
0.004
|
0.009
|
0.002
|
0.000
|
0.006
|
0.021
|
0.010
|
|
90°
|
0.040
|
0.012
|
0.064
|
0.007
|
0.000
|
0.016
|
0.139
|
0.029
|
|
120°
|
0.053
|
0.007
|
0.136
|
0.005
|
0.000
|
0.009
|
0.228
|
0.023
|
|
150°
|
0.014
|
0.003
|
0.023
|
0.003
|
0.000
|
0.003
|
0.039
|
0.015
|
|
180°
|
0.029
|
0.004
|
0.032
|
0.004
|
0.000
|
0.003
|
0.048
|
0.012
|
|
210°
|
0.074
|
0.007
|
0.083
|
0.004
|
0.000
|
0.004
|
0.100
|
0.014
|
|
240°
|
0.000
|
0.001
|
0.000
|
0.000
|
0.000
|
0.000
|
0.000
|
0.001
|
|
270°
|
0.000
|
0.000
|
0.000
|
0.000
|
0.000
|
0.000
|
0.000
|
0.001
|
|
300°
|
0.008
|
0.001
|
0.001
|
0.001
|
0.000
|
0.000
|
0.001
|
0.003
|
|
330°
|
0.043
|
0.004
|
0.040
|
0.003
|
0.000
|
0.004
|
0.027
|
0.007
|
Note on Atmospheric Stability
The Pasquill-Gifford atmosphere stability classes
range from A through F.
A: Turbulent
B: Very
unstable
C: Unstable
D: Neutral
E: Stable
F: Very
stable
Wind speed and solar radiation interact to determine
the level of atmospheric stability, which in turn suppresses or enhances the
vertical element of turbulent motion.
The latter is a function of the vertical temperature profile in the
atmosphere; the greater the rate of decrease in temperature with height, the
greater the level of turbulence.
Class A represents extremely unstable conditions,
which typically occur under conditions of strong daytime insolation. Category D is neutral and neither
enhances nor suppresses atmospheric turbulence. Class F on the other hand represents
moderately stable conditions, which typically arise on clear nights with little
wind.
The annual average temperature for Black Point is
23.9 °C.
Temperature data was not available from the Sha Chau station and so
temperature readings were taken from the Hong Kong Airport
instead. The average relative
humidity is 78%. Table 12B.7 below
tabulates temperature statistics.
Table 12B.7 Temperature
Statistics for Black Point
|
|
|
Min.
|
Max.
|
Average
|
|
Ambient air (T°C)1
|
BP
|
6.7
|
35.1
|
23.9
|
|
Surface (T°C)1
|
|
20.9
|
25.7
|
23
|
|
Seawater (T°C)2
|
BP
|
16.2
|
27.8
|
23.9
|
|
Humidity (%)1
|
|
65
|
82
|
77
|
Source: 1. Hong Kong
Observatory, “The Year’s Weather – 2003”
2. HK EPD, “Summary water quality
statistics of the Junk
Bay and Deep Bay WCZs in
2002”
12B.4
Hazard
identification
The hazard identification process is a formal review
to identify all hazards for the Gas Receiving station. This consisted of a review of the
hazardous properties of natural gas, a review of
accidents that have happened at similar facilities worldwide and a HAZID
workshop (see Section 12B.4.4). Hazards identified from these studies
are then carried forward for further consideration in the QRA.
For all hazards assessed as having a frequency of
less than 10-9 per year, the frequency assessment will be documented
but no quantification of consequences will be performed.
All scenarios with a frequency greater than 10-9
per year and potential to cause fatalities have the consequences of the event
quantified.
Hazard scenarios are excluded from the risk assessment
if one of the following conditions is satisfied:
·
The
frequency is below 1×10-9 per year.
·
The
frequency of a particular event is significantly smaller than other causes of
failure considered in the generic frequency.
·
If
the generic failure frequency is judged to include events of such kind, then
such events are not assessed separately.
·
If
there are no consequences. If an
event can be shown not to cause a loss of containment then the event is not
considered further.
12B.4.1
Hazards from Natural Gas
Hazards associated with natural gas (NG) have been
identified based on a review of known incident records worldwide and experience
gained from operations at similar facilities. The details are included below.
The main hazards associated with natural gas arise
from its flammability and the risk of fire. If NG is accidentally released, it will
mix with air to form a flammable mixture.
The plume will only ignite if it encounters an ignition source while
concentrated within its flammability range. In some cases, static discharges may
also cause immediate ignition of a release.
The characteristics of the possible hazardous effects
are described below.
Jet Fire
Jet fires result from ignited releases of pressurised
flammable gas. The
momentum of the release carries the materials forwards in a long plume
entraining air to give a flammable mixture. Jet fires only occur where the NG is
being handled under pressure. Since
the GRS will have NG at between 40 and 100 bar, jet fires are expected to be
the main hazard.
Fireball
Immediate ignition of releases caused by a rupture of
equipment/piping may give rise to a fireball upon ignition. Fireballs have very high thermal
radiation, similar to jet fires although the duration of the event is short.
Flash Fire
Following a NG release, if the cloud is not ignited
immediately, it will move with the wind and be diluted as a result of air
entrainment. The dispersing cloud
may subsequently come in contact with an ignition source and burn rapidly with
a sudden flash. Direct contact with
the burning gas may cause fatalities but the short duration of the flash fire
means that thermal radiation effects are not significant outside the cloud and
thus no fatalities are expected outside of the flash fire envelope.
Vapour Cloud Explosions
If a dispersing gas cloud accumulates in a confined
or congested area and is subsequently ignited, significant overpressures (an
explosion) may be generated. The
GRS, however, will be located in an open area without such confinement and an
unconfined cloud of natural gas is known not to produce damaging
overpressures. Vapour cloud
explosions are therefore not considered in this assessment; the worst effects
from a delayed ignition of a release will be a flash fire.
12B.4.2
Main Hazards from the Gas
Receiving Station
The main hazard from the GRS is loss of containment
from piping and equipment leading to a gas leak and fire.
Loss of Containment Incidents
The principal causes for loss of containment are:
·
corrosion
- internal and external;
·
third
party interference due to work in adjoining areas, etc;
·
material
defect;
·
construction
defect;
·
improper
operations;
·
defect
caused by pressure cycling; and
·
external - flooding, subsidence etc.
Review of Industry Incidents
A review of industry incidents at gas receiving
stations was carried out. Incident
records over the last few decades show releases and fires. These were associated with leaks from
valves and process equipment.
Based on the accident databases (MARS, ARIA) and
other information, some incidents examples are provided in Table 12B.8.
Table 12B.8 Incident
Review
|
Date, place
|
Cause
|
Description
|
Source
|
|
15/11/2007, USA
|
Unknown
|
An explosion occurred
at around 11.30 am in a natural gas treatment facility. It resulted in 4 injuries, two of them
severe.
|
ARIA
|
|
23/09/2002, USA
|
Unknown
|
In a natural
gas treatment facility, a flash fire like event occurred in the central part where
the raw natural gas is washed to remove impurities. Four of the nearby employees are
injured: 3 suffered severe burns and intoxication.
|
ARIA
|
|
28/12/2000, CANADA
|
Unknown
|
Explosion at a natural
gas pumping station rattled windows 1.5 miles away. There was no rupture of the pipeline
itself and the cause of the incident remains unknown. One man severely injured and gas
pressure to customers affected
|
MHIDAS
|
|
28/05/2000, CANADA
|
Overpressure
|
A 42” pipe
transporting natural gas ruptured during a pressure test. Authorities indicated that the gas
inlet was promptly shut down; environmental effects were therefore assumed to
be zero.
|
ARIA
|
|
04/01/1999, USA
|
Unknown
|
In a sub station
of a natural gas pipeline, a leakage led to an explosion and a fire
destroying a house and workshop.
The incident, visible from 30 km was taken care of by firemen and
controlled within 4 hours. Two
firemen suffered mild injuries.
|
ARIA
|
|
14/08/1998, USA
|
External events
|
Lightning
strike set fire to a natural gas compressor station. The resulting explosions sent a
fireball 600ft into the air. 5
people were injured. Gas supplies
to the whole of the Florida
peninsula were shut off. Residents
within 2 miles were evacuated.
|
MHIDAS
|
|
02/04/1998, RUSSIA
|
Unknown
|
The metering
unit of the natural gas distribution station was rocked by an explosion. A fire also occurred.
|
MHIDAS
|
|
25/06/2001, KAZAKHSTAN
|
Corrosion
|
Six metres of a
one metre diameter pipe was thrown 40 metres in the blast. Corrosion of the pipeline is thought
to have led to the leak that caused the blast. Fire quickly extinguished and supplies
resumed through an alternative pipe after three hours.
|
MHIDAS
|
|
10/04/2001, USA
|
Mechanical
failure
|
Residents were
evacuated for about three hours after a volatile gas cloud formed over a
natural gas facility. The source
of the leak was tracked down to a section of pipe, which was repaired.
|
MHIDAS
|
|
28/05/2000, CANADA
|
Mechanical
failure
|
A section of
the 42" pipeline ruptured during pressure-testing of the pipe.
|
MHIDAS
|
|
18/11/1998, UK
|
Impact
|
Workmen caused
a main gas pipeline to fracture, sending a 30ft plume of gas into the
air. Local residents were
evacuated and roads sealed off.
It was several hours before the pressure had dropped enough for the
pipe to be sealed off. No one was
injured.
|
MHIDAS
|
|
27/06/1997, USA
|
Human factor
|
Gas escaped from
a pipeline when equipment being used to take a metering station out of
commission fractured a valve. No
injuries were reported. People
within a mile of the rupture were evacuated. No fire or explosion occurred.
|
MHIDAS
|
|
08/02/1997, USA
|
Unknown
|
A leakage
occured on a natural gas pipeline of 660 mm diameter. The gas cloud exploded and a 100m high
flame occurred. Nearby houses
were shaken by the deflagration.
|
ARIA
|
|
01/01/1997, TURKEY
|
Human error
|
A natural gas
leak occured on a badly closed valve on a pipe (pressure= 20 bar). This incident led to death by
asphyxiation of the two employees who entered in the room, one equipped with
an inappropriate mask and the other without equipment
|
ARIA
|
|
18/12/1995, RUSSIA
|
Mechanical
failure
|
Section of
pipeline exploded due to high pressure in pipe.
|
MHIDAS
|
|
22/11/1995, RUSSIA
|
Corrosion
|
An explosion
followed by a fire occurred on a 0.5m diameter natural gas pipe. Corrosion is at the origin of the
accident. 240 m of pipes were
destroyed.
|
ARIA
|
|
19/03/1995, USA
|
Unknown
|
36" gas
pipe ruptured. Leak caught fire
& damaged reported 300ft section.
Gas rerouted to two parallel lines
|
MHIDAS
|
|
29/07/1993, UK
|
Impact
|
1000 workers were
evacuated as building contractors ruptured a mains pipe sending 40ft gas into
the air. Roads were sealed off
for about an hour while the leak was brought under control.
|
MHIDAS
|
|
18/05/1989, GERMANY
|
General
maintenance
|
Repairs to product
pipeline possibly caused explosions/fires which destroyed refinery
pumping/mixing station. Blaze
burned for 4hrs as fire fed by 100te of fuel leaking from broken pipe system.
|
MHIDAS
|
Pig Receiver Related Hazards
Operation of pig receivers poses a significant
hazard. There have been a few
incidents in the past relating to pig launcher/receiver operations. Usually it relates to the launcher and
receiver not being properly depressurised.
With pressure behind the pig, it launches like a cannon out the end of
the launcher or receiver when the operator opens the hatch.
Reasons that the receiver may not depressurise
properly include the following:
·
Vent/
drain valves are prone to plugging.
These are often small valves and can be plugged because of the dirty
material that has been scraped from the pipeline by the pig;
·
Stuck
pig trap. The pig could be stuck if
there is a restriction, probably a partially open valve, and there could be a trapped
pocket of high pressure gas;
·
Poor
pressure indication;
·
Human
errors, e.g. opening the trap door while the vessel is under pressure.
Pigging is generally performed infrequently. It is conservatively assumed in this
analysis that pigging operations will be performed once every 5 years. Therefore the risk contributed by the
pigging facilities to the whole system is relatively low. However, the risk per pigging operation
still remains high. Procedures and
devices are available to avoid accidents during pigging, and the human error is
therefore a major reason for pigging accidents. This issue is investigated further in
the frequency analysis, Section 12B.5.
Review of Industry Incidents
A review of industry incidents related to pig
facilities was carried out.
There are a number of well established international
accident databases that were considered for identification of hazards and
estimation of frequency of loss of containment incidents. The relevant accident databases are:
·
MHIDAS,
AEA Technology, UK [18]
·
IChemE
Accident Database [19]
·
MARS
(Major Accident Reporting System)
Based on the above accident databases and other
information, incidents associated with pigging are summarised below:
a)
At a
location in the Netherlands,
an uncontrolled release of approximately 10,000m3 of wet gas occurred from a pig receiver
at a drying facility due to the inadvertent opening of the pig receiver inlet
valve. This occurred due to a
malfunctioning motorised actuator that opened the receiver’s isolation valve
when the hinged door was not totally secure. (source : IChemE Accident database)
b)
Somewhere
in Western Europe, accidental closure by a
pipeline workman of a main line valve at a pump station caused a scraper pig
trap at an upstream facility to be over-pressurised. A spillage of 252 m3 of jet fuel occurred. The pipeline was out of service for two
days while the trap installation was modified. (source : IChemE Accident database)
c)
An
incident occurred in 2001 during pre-commissioning when a contractor was
dewatering a 10 mile section after a hydrostatic test. They were pushing a foam pig with air to
displace the water. The pig got
stuck somewhere in the pipe and they began pressuring up the section to
approximately 400 psig. The water
was being removed from a 12” bypass line.
They decided that the restriction was not allowing the pig to move
freely so they opened the end of the temporary trap. At this point, the pig had a downstream
pressure of ambient and upstream pressure of 400psig. In order to catch the pig, a large front
end loader was placed in front of the open trap. However, the pig shot out of the trap,
completely flipped the loader and continued to fly approximately 150 yards in
the air, destroying a wooden platform along the way.
d)
Two
workers were attempting to remove a pig from a line that was launched the
previous day. They found the pig was
stuck in the reducer section but cleared the block valve so the valve could be
closed. After depressurising pig
receiver and opening it to atmosphere, both workers believed the pig can be
removed. The pig had to be pulled
into the receiver through the reducer to remove it so they fashioned a hook
from some SS tubing. When worker
hooked onto pig, the pig shot out and struck him in the face resulting in major
injuries. Apparently, part of the
pig had created a seal with a weld in the reducer section, which created a
pressure trap behind the pig.
(source : www.offshoreman.com)
e)
There
is also an incident reported to have occurred in the North
Sea offshore when the pig barrel cover was blown away and the pig
flew about 1.5km into the sea. The
incident apparently occurred due to misoperation of the valves on the pig
barrel leading to sudden impact of the pig on the barrel cover.
f)
During
a pigging operation, a 30" pipeline, bringing natural gas on-shore, failed
between the emergency shut-down (ESD) valve and the pig trap. The intervention of the ESD valve from
the control room failed and it had to be closed locally. Released natural gas did not ignite, but
was a serious risk to personnel.
(source: MARS database)
The above list of incidents shows the potential for
misoperation leading to the barrel cover or the pig shooting out of the trap
and causing damage to nearby equipment and buildings in addition to causing
operator injury.
Vent Stack Related Hazards
A common vent stack will be provided for emergency
depressurization of the existing and new GRS facilities. The vent will be designed according to
appropriate standards such as API 521.
This ensures that the concentration of gases at ground level or on any
elevated structures will not reach the lower flammability limit. Also, in the event of an ignited
release, thermal radiation at ground level and on elevated structures will meet
the standards. As such, there is no
consequence from emergency venting and the vent stack is not considered further
in the analysis.
12B.4.3
External
and Natural Hazards
Seismic Hazard
An earthquake has the potential to cause damage to
pipework and process vessels.
Damage to pipework could be due to ground movement/vibration, with guillotine
failure of pipes.
Studies by the Geotechnical Engineering Office
conducted in the last decades indicate that Hong Kong SAR is a region of low
seismicity (e.g., GCO, 1991 [8]; GEO, 2002-2004 [9]). The
seismicity in the vicinity of Hong Kong is considered similar to that of areas
of Central Europe and the Eastern areas of the USA [10]. As Hong Kong is a region of low seismicity, an earthquake is
an unlikely event. The generic
failure frequencies adopted in this study are based on historical incidents
that include earthquakes in their cause of failure. Since Hong Kong
is not at disproportionate risk from earthquakes compared to similar facilities
worldwide, it is deemed appropriate to use these generic frequencies without
adjustment. There is no need to
address earthquakes separately as they are already included in the generic
failure rates.
Subsidence
For subsidence which would result in failure of
pipework or vessels, the ground movement must be relatively sudden and
severe. Normal subsidence events
occur gradually over a period of months and thus appropriate mitigating action
can be taken to prevent failures. In
the worst cases, the plant would be shut down and the relevant equipment
isolated and depressurised. The GRS
will be built on coastal land with solid foundation. No undue risk from subsidence is
therefore expected and failures due to this are deemed to be included in
generic failure frequencies.
Lightning
Lightning strikes have led to a number of major
accidents world-wide. For example,
a contributory cause towards the major fire at the Texaco refinery in the UK in 1994 was
thought to be an initial lightning strike on process pipework.
The installation will be protected with lightning
conductors to safety earth direct lightning strikes. The grounding will be inspected
regularly. The potential for a lightning
strike to hit the facility and cause a release event is therefore deemed to be
unlikely. Failures due to lightning
strikes are taken to be covered by generic frequencies.
Storm Surges and Flooding
If the piping become
submerged under water, it is possible for buoyancy forces to lift the
pipes/tanks, causing damage and possible loss of containment.
Flooding from heavy
rainfall is not possible due to the coastal location of the site. The slopes of the natural terrain will
channel water to the sea. The
primary hazard from typhoons is the storm surge. Winds, and to a lesser extent pressure,
cause a rise in sea level in coastal areas. In general, storm surges are limited to
several metres.
The GRS facilities,
located +6mPD above sea level are therefore protected against any risk from
storm surges, waves and other causes of flooding.
Tsunami
Similar to storm surges,
the main hazard from tsunamis is the rise in sea level and possible floatation
of piping and tanks. The highest
rise in sea level ever recorded in Hong Kong due to a tsunami was 0.3m [15], and occurred as a result of the 1960 earthquake in Chile, the
largest earthquake ever recorded in history at magnitude 9.5 on the Richter
scale. The GRS site is
approximately at +6mPD. The effect
of a tsunami on the GRS is therefore considered negligible.
The reason for the low
impact of tsunamis on Hong Kong may be explained by the extended continental
shelf in the South China Sea which effectively
dissipates the energy of a tsunami.
Also, the presence of the Philippine Islands and Taiwan act as
an effective barrier against seismic activity in the Pacific [16].
Secondary waves that pass through the Luzon
Strait diffract and lose energy as
they traverse the South China Sea.
Seismic activity with the South China Sea area may also produce tsunamis. Earthquakes on the western coast of
Luzon in the Philippines
have produced localised tsunamis but there is no record of any observable
effects in Hong Kong.
Summary of Natural Hazards
The GRS site and design of the facility are such that
there will be no special risks from natural hazards. Natural hazards are therefore not
treated separately in the analysis but are included in the generic failure
frequencies.
Aircraft Crash
The Black Point site does not lie within the flight
path of Chek Lap Kok (Figure 12B.8),
being about 10km from the nearest runway.
Figure 12B.8 Flight
Paths at Hong Kong
International Airport
The frequency of aircraft crash was estimated using
the methodology of the HSE [11]. The
model takes into account specific factors such as the target area of the
proposed hazard site and its longitudinal (x)
and perpendicular (y) distances from
the runway threshold (Figure 12B.9). The crash frequency per unit ground area
(per km2) is calculated as:
(4)
Where N is
the number of runway movements per year and R
is the probability of an accident per movement (landing or take-off). F(x,y) gives the spatial distribution of crashes and is given
by:
Landings
(5)
for 
km
Take-off
(6)
for 
km
Equations 5 and 6
are valid only for the specified range of x
values. If x lies outside this range, the impact probability is zero.
Figure 12B.9 Aircraft
Crash Coordinate System
NTSB data [12] for fatal accidents in the U.S. involving scheduled
airline flights during the period 1986-2005 are given in Table 12B.9. The
10-year moving average suggests a downward trend with recent years showing a
rate of about 2×10-7 per flight. However, only 13.5% of accidents are
associated with the approach to landing, 15.8% are associated with take-off and
4.2% are related to the climb phase of the flight [13]. The
accident frequency for the approach to landings hence becomes 2.7×10-8
per flight and for take-off/climb 4.0×10-8 per flight. The number of flights at Chep Lap Kok
for year 2011 is conservatively estimated at 394,000 (a 50% increase over
2005).
Table 12B.9 U.S
Scheduled Airline Accident Rate [12]
|
Year
|
Accident
rate per 1,000,000 flights for accidents involving fatalities
|
10-year
moving average accident rate per 1,000,000 flights
|
|
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
|
0.14
0.41
0.27
1.10
0.77
0.53
0.53
0.13
0.51
0.12
0.38
0.30
0.09
0.18
0.18
0.19
0.00
0.2
0.09
0.27
|
-
-
-
-
-
-
-
-
|
0.451
|
|
0.475
|
|
0.464
|
|
0.446
|
|
0.354
|
|
0.295
|
|
0.261
|
|
0.208
|
|
0.215
|
|
0.173
|
|
0.188
|
|
Considering landings on runway 25R for example, the
values for x and y according to Figure 12B.8
are 0.6 and 10.7km respectively.
Applying Equation 5 gives FL=
8.1×10-5 km-2.
Substituting this into Equation 4
gives:
/year/km2
The number of plane movements has been divided by 8
to take into account that half of movements are take-offs and only a quarter of
landings use runway 07R. This
effectively assumes that each runway is used equally.
The target area is estimated at 12,000m2
or 0.012km2. This gives
a frequency for crashes on the site associated with landings on runway 07R as
2.2×10-9 per year.
Repeating the calculation for landings and take-offs from all runways
gives the results shown in Table 12B.10.
Table 12B.10 Aircraft
Crash Frequency onto the GRS
|
Runway
|
Landing (per year)
|
Take-off (per year)
|
|
07R
07L
25L
25R
|
0
0
1.3×10-9
8.6×10-9
|
1.3×10-11
2.6×10-12
0
0
|
|
Total
|
2.2×10-9
|
1.6×10-11
|
The combined frequency of all take-off and
landing crashes onto the GRS from activities on all runways is less than 2.2×10-9
per year. This frequency is small
compared to the generic frequencies used in the study. Aircraft crash is therefore neglected from
the analysis.
Helicopter Crash
Helipad Activity
The Black Point Power Station site is provided with a
helicopter landing pad although the frequency of use is expected to be low with
perhaps one landing/take-off per week.
The approach, landing and take-off stages of an aircraft flight are
associated with the highest risk and therefore the possible impact of
helicopter crashes on the facility were assessed.
Data from offshore helicopter activities [14] gives a helipad related helicopter crash frequency of
2.9×10-6 per flight stage (i.e. per take-off and landing). However, most of these incidents are
minor such as heavy landings. For a
helicopter incident to damage the facility, it must be a serious, uncontrolled
impact. Only accidents involving
fatalities were therefore considered in the analysis. 4% of incidents resulted in one or more
fatalities and so the frequency of uncontrolled crashes was calculated as
2.9×10-6 × 0.04 = 1.2×10-7 per flight stage. For one flight per week using the
helipad, the annual crash frequency becomes 1.2×10-7 × 52 = 6.0×10-6
/year.
Helicopter accidents during take-off and landing are
confined to a small area around the helipad [11]. 93% of
accidents occur within 100m of the helipad. The remaining 7% occur between 100 and
200m of the helipad. There have
been no serious helipad related incidents resulting in a crash beyond 200m of
the helipad.
The distance of Gas Receiving Station to the helipad
is more than 400 metres. Helicopter
crash is therefore not considered further in this study.
Passing Helicopters
There are no helicopter flight paths near BPPS. The possibility of a passing helicopter
crashing into the GRS facility is therefore much smaller than the generic
failure frequencies used in this study.
Helicopter crashes are therefore not considered separately but are
deemed to be included in the generic failure frequencies.
12B.4.4
HAZID Study
A Hazard Identification (HAZID) Study was conducted
in September 2009. The potential
hazards posed by the facility were identified based on the HAZID team’s expert
opinion, past accidents, lessons learnt and checklists. The details of the HAZID study can be
found in Table 12B.11.
A systematic approach was adopted, whereby the facility was divided into a
number of “subsystems” based on the layout and the process; guidewords from the
checklist were then applied to each
subsystem as relevant.
Figure 12B.12 Misoperation to Open Pig Barrel Door
Figure 12B.13 Misoperation Leading
to Pig Impact on the Barrel Cover
12B.5.4
Construction Activities
The GRS construction may present an
increase in risk due to construction activities from the GRS impacting on
existing facilities.
The project has taken this into
consideration with the following safeguards:
·
Safety
management system and procedures will be developed for the new GRSs. Details of the system and procedures
will be given in the safety case study for the new GRSs. Systems relating to construction
activities, such as Fire and Safe Work Permit System, risk assessment, and
emergency response procedure, will be in place and enforced before commencement
of work. Recommendations in
accordance with best practice have also been given to protect the workers at
the sites (see section 12.3.2);
·
Good
access is provided to construction areas with access roads from at least 2
sides of the site;
·
The
reclamation itself will be formed mostly by accessing from the sea.
The Gas Production & Supply Code of
Practice [17] provides a practical guidance in respect of the
requirements of the Gas Safety Ordinance Cap 51 and the Gas Safety (Gas Supply)
Regulations. Article 23A of these
regulations requires that:
·
No
person shall carry out, or permit to be carried out, any works in the vicinity
of a gas pipe unless he or the person carrying out the works has before
commencing the works, taken all reasonable steps to ascertain the location and
position of the gas pipe; and
·
A
person who carries out, or who permits to be carried out, any works in the
vicinity of a gas pipe shall ensure that all reasonable measures are taken to
protect the gas pipe from damage arising out of the works that would be likely
to prejudice safety.
Work, ‘in the vicinity’ of gas pipes is
defined according to Table 12B.16. Although many of the activities listed
are not directly relevant to the GRS, Table
12B.16 serves to indicate typical effects distances for different types of
work and when special precautions are warranted. The GRS separation distances mostly
exceed those listed in Table 12B.16. However, some of the minimum separation
distances specified in Table 12B.16
are not met. These activities
include but not limited to the following:
·
Construction
of the pre-heaters and pressure reduction station for Phase 1 close to the
existing underground high pressure gas pipe line (~10 m)
·
Construction
of new pipe racks adjacent to the existing pipe rack
·
Construction
of the inlet pipeline for Second Phase close to the existing pre-heaters (~20
m)
Special consideration may be given to the
underground high pressure pipe line since the external load on the ground by
the crane and the drilling operation may pose additional risk of damage. However, with the implementation of
safety procedures, the risk is not expected to be significantly higher than the
generic failure frequencies adopted. Nevertheless, a Job Safety Study will be
conducted to assess the potential risk and failure modes of such construction
operations and special precautions will be included in the procedure.
Based on the above, the likelihood of
damage to the operational facilities from construction activities will be
low. This is therefore not
considered further in this study.
Table
12B.16 Works
in the Vicinity of Gas Pipes
|
Type of Work
|
Distance
|
|
Trench or other excavation up to 1.5m in depth in
stable ground
|
10m
|
|
Trench or other excavation over 1.5m and up to 5m in
depth
|
15m
|
|
Trench or other excavation in stable ground over 5m
in depth
|
20m
|
|
Welding or hot works near exposed gas pipes or
above ground installations
|
10m
|
|
Piling,
percussion moling or pipe bursting
|
15m
|
|
Works
near high pressure pipelines
|
20m
|
|
Ground
investigation and any kind of drilling or core sampling
|
30m
|
|
Use
of explosives
|
60m
|
12B.6
Consequence
Analysis
12B.6.1
Source Term Modelling
The process facility was divided into 9 isolatable
sections. Table 12B.12 listed the process details adopted for each process
section. Discharge rates for each
given leak size were modelled using standard orifice type calculations
contained within the PHAST suite of models. Design pressures were used as a conservative
approach in the consequence modelling.
12B.6.2
Consequence Modelling
Table 12B.17 shows the list of release scenarios along
with the corresponding consequence model used in PHAST.
Table
12B.17 Release
Scenarios and Consequence Models Applied
|
Release
Scenario
|
Release Type
|
Model
Applied in PHAST
|
|
10mm leak
|
Leak
|
Leak
|
|
25mm leak
|
Leak
|
Leak
|
|
50mm leak
|
Leak
|
Leak
|
|
100mm leak
|
Leak
|
Leak
|
|
Full bore
rupture
|
Rupture
|
Catastrophic Rupture
|
The consequence modelling parameters for PHAST are
listed in Table 12B.18.
Table
12B.18 Consequence
Modelling Parameters
|
BLEVE Parameters
|
|
|
|
|
|
Maximum SEP for a BLEVE
|
|
400.00
|
kW/m2
|
|
|
Fireball radiation intensity level 1
|
7.00
|
kW/m2
|
|
|
Fireball radiation intensity level 2
|
14.00
|
kW/m2
|
|
|
Fireball radiation intensity level 3
|
21.00
|
kW/m2
|
|
|
Mass Modification Factor
|
|
3.00
|
|
|
|
Fireball Maximum Exposure Duration
|
30.00
|
s
|
|
|
Ground Reflection
|
|
Ground Burst
|
|
|
|
Ideal Gas Modelling
|
|
Model as real gas
|
|
|
|
|
|
|
|
|
Discharge Parameters
|
|
|
|
|
|
Continuous Critical Weber number
|
12.50
|
|
|
|
Instantaneous Critical Weber number
|
12.50
|
|
|
|
Venting equation constant
|
|
24.82
|
|
|
|
Relief valve safety factor
|
|
1.20
|
|
|
|
Minimum RV diameter ratio
|
|
1.00
|
|
|
|
Critical pressure greater than flow phase
|
0.34
|
bar
|
|
|
Maximum release velocity
|
|
500.00
|
m/s
|
|
|
Minimum drop size allowed
|
|
0.00
|
mm
|
|
|
Maximum drop size allowed
|
|
10.00
|
mm
|
|
|
Default Liquid Fraction
|
|
1.00
|
fraction
|
|
|
Continuous Drop Slip factor
|
|
1.00
|
|
|
|
Instantaneous Drop Slip factor
|
|
1.00
|
|
|
|
Pipe-Fluid Thermal Coupling
|
|
0.00
|
|
|
|
Number of Time Steps
|
|
100.00
|
|
|
|
Maximum Number of Data Points
|
|
1,000.00
|
|
|
|
Non-Return Valve velocity head losses
|
0.00
|
|
|
|
Pipe roughness
|
|
0.046
|
mm
|
|
|
Shut-Off Valve velocity head losses
|
0.00
|
|
|
|
Excess Flow Valve velocity head losses
|
0.00
|
|
|
|
Default volume changes
|
|
3.00
|
/hr
|
|
|
Line length
|
|
10.00
|
m
|
|
|
Elevation
|
|
1.00
|
m
|
|
|
Atmospheric Expansion Method
|
|
Closest to Initial Conditions
|
|
|
|
Tank Roof Failure Model Effects
|
|
Instantaneous Effects
|
|
|
|
Outdoor Release Direction
|
|
Horizontal
|
|
|
|
|
|
|
|
|
Dispersion Parameters
|
|
|
|
|
|
Dense cloud parameter gamma (continuous)
|
0.00
|
|
|
|
Dense cloud parameter gamma (instant)
|
0.30
|
|
|
|
Dense cloud parameter k (continuous)
|
1.15
|
|
|
|
Dense cloud parameter k (instantaneous)
|
1.15
|
|
|
|
Jet entrainment coefficient alpha1
|
0.17
|
|
|
|
Jet entrainment coefficient alpha2
|
0.35
|
|
|
|
Ratio instantaneous/continuous sigma-y
|
1.00
|
|
|
|
Ratio instantaneous/continuous sigma-z
|
1.00
|
|
|
|
Distance multiple for full passive entrainment
|
2.00
|
|
|
|
Quasi-instantaneous transition parameter
|
0.80
|
|
|
|
Impact parameter - plume/ground
|
0.80
|
|
|
|
Expansion zone length/source diameter ratio
|
0.01
|
|
|
|
Drop/expansion velocity for inst.
release
|
0.80
|
|
|
|
Drag coefficient between plume and ground
|
1.50
|
|
|
|
Drag coefficient between plume and air
|
0.00
|
|
|
|
Default bund height
|
|
0.00
|
m
|
|
|
Maximum temperature allowed
|
|
626.85
|
degC
|
|
|
Minimum temperature allowed
|
|
-263.15
|
degC
|
|
|
Minimum release velocity for cont. release
|
0.10
|
m/s
|
|
|
Minimum integration step size (Instantaneous)
|
0.10
|
s
|
|
|
Maximum integration step size (Instantaneous)
|
1,000.00
|
s
|
|
|
Minimum integration step size (Continuous)
|
0.10
|
m
|
|
|
Maximum integration step size (Continuous)
|
100.00
|
m
|
|
|
Maximum distance for dispersion
|
50,000.00
|
m
|
|
|
Maximum height for dispersion
|
|
1,000.00
|
m
|
|
|
Minimum cloud depth
|
|
0.02
|
m
|
|
|
Expansion energy cutoff for droplet angle
|
0.69
|
kJ/kg
|
|
|
Droplet evaporation thermodynamics model
|
Rainout, Non-equilibrium
|
|
|
|
Flag for mixing height
|
|
Constrained
|
|
|
|
Accuracy for integration of dispersion
|
0.00
|
|
|
|
Accuracy for droplet integration
|
|
0.00
|
|
|
|
Richardson number
criterion for cloud lift-off
|
-20.00
|
|
|
|
Flag to reset rainout position
|
|
Do not reset rainout position
|
|
|
|
Surface over which the dispersion occurs
|
Water
|
|
|
|
Minimum Vapor Fraction for Convection
|
0.00
|
fraction
|
|
|
Coefficient of Initial Rainout
|
|
0.00
|
|
|
|
Minimum Continuous Release Height
|
0.00
|
m
|
|
|
Flag for finite duration correction
|
Finite Duration Correction
|
|
|
|
Near Field Passive Entrainment Parameter
|
1.00
|
|
|
|
Jet Model
|
|
Morton et.al.
|
|
|
|
Maximum Cloud/Ambient Velocity Difference
|
0.10
|
|
|
|
Maximum Cloud/Ambient Density Difference
|
0.02
|
|
|
|
Maximum Non-passive entrainment fraction
|
0.30
|
|
|
|
Maximum Richardson
number
|
|
15.00
|
|
|
|
Core Averaging Time
|
|
18.75
|
s
|
|
|
Ground Drag Model
|
|
New (Recommended)
|
|
|
|
Flag for Heat/Water vapor transfer
|
Heat and Water
|
|
|
|
Richardson Number for passive transition above pool
|
0.02
|
|
|
|
Pool Vaporization entrainment parameter
|
1.50
|
|
|
|
Modeling of instantaneous expansion
|
Standard Method
|
|
|
|
Minimum concentration of interest
|
0.00
|
fraction
|
|
|
Maximum distance of interest
|
|
10,000.00
|
m
|
|
|
Model In Use
|
|
Best Estimate
|
|
|
|
Maximum Initial Step Size
|
|
10.00
|
m
|
|
|
Minimum Number of Steps per Zone
|
5.00
|
|
|
|
Factor for Step Increase
|
|
1.20
|
|
|
|
Maximum Number of Output Steps
|
1,000.00
|
|
|
|
|
|
|
|
|
Flammables Parameters
|
|
|
|
|
|
Height for calculation of flammable effects
|
0.00
|
m
|
|
|
Flammable result grid step in X-direction
|
10.00
|
m
|
|
|
LFL fraction to finish
|
|
0.85
|
|
|
|
Flammable angle of inclination
|
|
0.00
|
deg
|
|
|
Flammable inclination
|
|
Variable
|
|
|
|
Flammable mass calculation method
|
Mass between LFL and UFL
|
|
|
|
Flammable Base averaging time
|
|
18.75
|
s
|
|
|
Cut Off Time for Short Continuous Releases
|
20.00
|
s
|
|
|
Observer type radiation modelling flag
|
Planar
|
|
|
|
Probit A Value
|
|
-36.38
|
|
|
|
Probit B Value
|
|
2.56
|
|
|
|
Probit N Value
|
|
1.33
|
|
|
|
Height for reports
|
|
Centreline Height
|
|
|
|
Angle of orientation
|
|
0.00
|
deg
|
|
|
Relative tolerance for radiation calculations
|
0.02
|
fraction
|
|
|
|
|
|
|
|
General Parameters
|
|
|
|
|
|
Maximum release duration
|
|
3,600.00
|
s
|
|
|
Height for concentration output
|
|
0.00
|
m
|
|
|
|
|
|
|
|
Jet Fire Parameters
|
|
|
|
|
|
Maximum SEP for a Jet Fire
|
|
400.00
|
kW/m2
|
|
|
Jet Fire Averaging Time
|
|
20.00
|
s
|
|
|
Jet fire radiation intensity level 1
|
|
7.00
|
kW/m2
|
|
|
Jet fire radiation intensity level 2
|
|
14.00
|
kW/m2
|
|
|
Jet fire radiation intensity level 3
|
|
21.00
|
kW/m2
|
|
|
Rate Modification Factor
|
|
3.00
|
|
|
|
Jet Fire Maximum Exposure Duration
|
30.00
|
s
|
|
|
Model Correlation Type
|
|
Shell
|
|
|
|
|
|
|
|
|
Weather Parameters
|
|
|
|
|
|
Atmospheric pressure
|
|
1.01
|
bar
|
|
|
Atmospheric molecular weight
|
|
28.97
|
|
|
|
Atmospheric specific heat at constant pressure
|
1.00
|
kJ/kg.degK
|
|
|
Wind speed reference height (m)
|
|
10.00
|
m
|
|
|
Temperature reference height (m)
|
0.00
|
m
|
|
|
Cut-off height for wind speed profile (m)
|
1.00
|
m
|
|
|
Wind speed profile
|
|
Power Law
|
|
|
|
Atmospheric Temperature and Pressure Profile
|
Temp.Logarithmic; Pres.Linear
|
|
|
|
Atmospheric temperature
|
|
23.00
|
degC
|
|
|
Relative humidity
|
|
0.77
|
fraction
|
|
|
Surface Roughness Parameter
|
|
0.043
|
|
|
|
Surface Roughness Length
|
|
0.912
|
mm
|
|
|
Roughness or Parameter
|
|
Parameter
|
|
|
|
Dispersing surface temperature
|
|
23.00
|
degC
|
|
|
Default surface temperature of bund
|
23.00
|
degC
|
|
|
Solar radiation flux
|
|
0.50
|
kW/m2
|
|
|
Building Exchange Rate
|
|
4.00
|
/hr
|
|
|
Tail Time
|
|
1,800.00
|
s
|
12B.6.3
Consequence End-Point Criteria
The end-point criteria are
used to define the impact level at which a fatality could result.
Thermal Radiation
The following probit equation [1] is used to determine impacts of thermal radiation
from jet fires to persons unprotected by clothing.
Y = -36.38 + 2.56 ln (t I 4/3) (1)
where I
is the radiant thermal flux (W/m2) and Y is the probit function which is related to the probability of
fatality. This equation gives the
data points presented in Table
12B.19, assuming a 30-second exposure time. For areas lying between any two
radiation flux contours, the equivalent fatality level
is estimated as follows:
·
For areas beyond the 50% fatality contour, the
equivalent fatality is calculated using a 2/3 weighting towards the lower
contour. For example, the
equivalent fatality between the 1% and 50% contours is calculated as 2/3 x 1 +
1/3 x 50 = 17%;
·
For areas within the 50% contour, the equivalent
fatality is calculated with a 2/3 weighting towards the upper contour. For example, the equivalent fatality
between the 90% and 50% contours is calculated as 2/3 x 90 + 1/3 x 50 = 77%.
The different approach
above and below the 50% fatality contour is due to the sigmoid shape of the
probit function.
Table
12B.19 Levels
of Harm for 30s Exposure to Heat Fluxes
|
Incident
Thermal Flux (kW/m2)
|
Fatality Probability
for 30s Exposure
|
Equivalent
Fatality Probability for Area between Radiation Flux Contours
|
|
7.3
|
1%
|
}
}
}
|
17%
|
|
14.4
|
50%
|
77%
|
|
20.9
35.5
|
90%
99.9%
|
97%
|
Fireballs are modelled in a similar manner as jet fires,
using the same probit equation.
However, fireballs are generally of shorter duration than 30 seconds and
hence the actual duration of the fireball was used to determine harm
probabilities.
Flash Fire
With regard to flash fires, the criterion chosen is
that a 100% fatality is assumed for any person outdoors within the flash fire
envelope. In this study, the extent
of the flash fire is assumed to be the dispersion distance to 85% of the LFL
for a conservative evaluation.
12B.6.4
Consequence Results
A complete list of hazard distances obtained from the
consequence modelling is provided in Table
12B.20.
