A Holistic Risk Assessment Method for Storage Tank

Terminals Considering Domino Effects Triggered by

Natural Hazards

AL I S AR I , P H . D . , P. E .
D I R E C T O R , O M E G A A D V A N C E D E N G I N E E R I N G A N D R I S K C O N S U LT I N G , “ O M E G A - R I S K ” , H O U S T O N , T X , U S A

A S S O C . P R O F. , I S T A N B U L T E C H N I C A L U N I V E R S I T Y, I S T A N B U L , T U R K E Y

N AT E C H R I S K M A N A G E M E N T W O R K S H O P 2 0 1 8

10 October 2018 1
 Presentation Outline
 Objectives
 Process and Natural Hazards at Storage Tank Terminals
 Risk Calculation Approach
 Risk Criteria
 Random Variables and Input Data for Risk Study
 Frequency Assessment
 Structural Consequence Assessment/ Structural Damage
 Examples - Hazard Consequence and Individual Risk
 Domino Effects
 Conclusions

2
Objectives
Risk assessment for a hydrocarbon storage tank terminal requires due consideration of process and
non-process hazards;
To date, there is no a comprehensive approach, or a set of industry guidelines, or technical
publications available that address the risks from hazards such as fire, blast, toxic smoke, tornado,
lightning, earthquake, loss of containment (dike failure, etc.), forest fires and etc.;
A significant gap in the industry that lacks to combine the knowledge of process safety, advanced
structural analysis, and reliability together to accurately and reliably estimate the individual,
environmental and facility damage risks from such hazards.
Lastly, there is a considerable need for inclusion of domino (knock-on) effects in the analysis
whereby multiple failures and catastrophic events are initiated simultaneously or in very close
proximity to each other.
Conclusion: there is a need for a comprehensive, holistic approach for determining risk in and around
the hydrocarbon storage tank terminal accounting for domino effects.

3
 Potential Process and Non-Process Hazards
PRIMARY EVENTS ESCALATIONS
(Secondary, Tertiary .. Events)
• Fire (pool fire, rim seal fire, pontoon fire,
Process roof spill fire, tank full surface fire, bund
fire and jet fire) - Boil-over
Hazards • Vapor Cloud Explosion
• Toxic*
- Dyke (Bund) Fire
- Full Surface Fire
(Single or Multiple Tank)
• Lightning
Natural • Seismic Event - Toxic Smoke
• Tornado - Loss of Containment
Hazards • Flooding/Rainstorm/Hurricane
- Forest Fire
- Explosion
• Terrorist Attack
Man – Made • Fire – Fighting /Tank Cooling Activities**

Hazards
4
 Potential Hazards at Storage Tank Facilities: Containment Loss
Reasons of Containment Loss:
1.Tank Rupture due to fire, blast, tornado and seismic
2. Rainstorm water
3. Firefighting activities

Fire Blast Seismic Tornado

Caribbean Petroleum Tank Terminal, Buncefield Hertfordshire Oil Storage Depot after Blast and Tank farm fires after Kocaeli Earthquake in Flipped Oil Tank after Tornado
2009 Puerto Rico Fire Stormed Tank Farm Turkey in 1999 in Edmonton in 1987
Tank Rupture: Oil discharged
Rainstorm
Fire-fighting

Containment Loss
The 2017 Arkema plant explosion during Hurricane Harvey
in Crosby, Texas
 Risk Assessment Flow Diagram
Tank and Process Information
P&IDs, Flow Diagrams, Fire-Fighting system
Shutdown Systems, Tank Utilization

Major Hazard Identification

Hazardous Materials, Failure Modes,
Loss Scenarios

Frequency Assessment Consequence Assessment

Toxic smoke, blast load, fire dimensions, PGA,
Fire/Blast, Toxic, Seismic, Tornado, rainstorm, ..
thermal radiation, structural damage, loss of
containment , domino effects
Human response to blast,
thermal effects, toxic smoke Risk Calculations Stop If
etc. Individual / Societal / Risk is
Environmental Risk Tolerable
Damage to environment due to loss
of containment
If Risk Not Tolerable
Identify Mitigation Measures / Re Analyze and Re - Calculate
Risk Until Tolerability/Acceptability Criteria Met
6
 Risk Calculation

• Risk due to hazards at a storage tank terminal and its surroundings is composed of summation of all risks
given no escalation (i.e. no domino effects) of undesired events and all risks given an escalation (i.e. domino
effects) of undesired events:

𝑅𝑖𝑠𝑘 = 𝑅𝑖𝑠𝑘 | 𝑁𝑜 𝐸𝑠𝑐𝑎𝑙𝑎𝑡𝑖𝑜𝑛 + 𝑅𝑖𝑠𝑘 | 𝐸𝑠𝑐𝑎𝑙𝑎𝑡𝑖𝑜𝑛

7
Risk Criteria – Major Industrial Accidents Council of Canada (MIACC)

Acceptable Levels of Risk on Annual

Individual Risk Basis 8
Some of the Input Data and Random Variables

 Wind data (wind rose, direction, speed and associated probability);

 Tank data for each tank (utilization probability, diameter, height, roof type, material);
 Fire-fighting and tank cooling fluid volumes;
 Site information (layout, coordinates, secondary/tertiary containment areas, shared
containment etc.);
 Hurricane, tornado hazard intensities/frequencies, paths, etc.;
 Peak Ground Accelerations (Seismic hazard curves);
 Tank content properties
 Emergency response time and probability to put out a fire
 Rainfall intensities, duration, and frequencies

9
 Analysis Random Variables
Example Tank Utilization
0.0% 4.0% 8.0% 12.0% 16.0% 20.0% 24.0% 28.0% 32.0% 36.0%
Histogram
40.0% 44.0% 48.0% 52.0% 56.0% 60.0% 64.0% 68.0% 72.0% 76.0% 80.0% 84.0% 88.0% 92.0% 96.0%
100% 100%
Frequency
90% Cumulative F requency
90%

80% 80%

Cumulative Frequency
70% 70%
Frequency

60% 60%

50% 50%

40% 40%

30% 30%

20% 20%

10% 10%

0% 0%
0 1 2 3 4 5 7 8 9 10 11 12 13 14 15 16 17 19 20 21 22 23 24 25 26 27
Tank Contents (kbbl)

Example Wind Rose: Speed and

Direction Probability

10
Analysis Input Data – Rain Precipitation

4.8 mm
100 yr
10 yr
3.4 mm

24 hr
Sumas Terminal

5 km

Source for
data collection 11
 Seismic Hazard Curves – PGA vs frequency
• Site seismic hazard curves can be developed for a specific site; or
• Can also be obtained from the National Seismic Codes and modified based on local soil characteristics.

http://www.earthquakescanada.nrcan.gc.ca/hazard-alea/interpolat/index_2015-eng.php

12
Tornado Frequency Map

• Tanks Location in the Tornado Frequency Map

Ref: Cheng, V.Y.S., Arhonditsis, G.B., Sills, D.M., Auld, H., Shephard, M.W., Gough, W.A., Klaasen, J.,
“Probability of Tornado Occurrence across Canada”, American Meteorological Society, December 13
2013. DOI: 10.1175/JCLI-D-13-00093.1
 Example - Calculations of Loss of Containment
• Probability of overfill due to Earthquake causing rupture and release:

F (OF EQ )  F ( EQ )  P(utilizatio n)  P( rain)  P( D EQ )

Frequency of overfill due to Frequency of Probability of damage to a tank due

earthquake earthquake to earthquake

14
 Modeling Procedure – Example Steel Cone Roof Tank
Structural Damage due to Blast
Finite element model of the tank includes:

• Tank
• Crude oil inside the tank
• Soil underneath the tank

Tank

1.2
θ=0 θ=10
Crude 1
θ=20 θ=30
θ=40 θ=50
Oil 0.8
θ=60 θ=70
θ=80 θ=90
θ=100 θ=110

P/Pr
0.6
θ=120 θ=130
θ=140 θ=150
0.4
θ=160 θ=170
θ=180
0.2

Soil 0
0 0.1 0.2 0.3 0.4 0.5
-0.2
Time
15
 Blast Response of Tank (Example Case)
Deformation response (m): Ps=3.5 psi (24 kPa), td=0.5 sec

16
 Structural Damage due to Fire
The modeling procedure includes following steps:

Step 1: Carry out a coupled heat transfer-CFD analysis to quantify the temperature evolution over the fire
exposed side of the tank.
Step 2: Carry out structural analysis to predict the performance of the tank at elevated temperature.

Thermal Analysis Structural Analysis

17
 Results for Structural Analysis – External Floating Roof
Time to failure (ttf) of open roof tanks based on yield criterion:
1000 Utilization: 25% 1000
Utilization: 50%
Landucci method
Landucci method
FEM
FEM

Time to Failure (sec)

Time to Failure

100 100

10 10
0 50 100 150 200 0 50 100 150 200
Heat flux (kW/m2) Heat flux (kW/m2)

1000
Utilization: 75% Landucci method

FEM_yield
Time to Failure (sec)

100

10
0 50 100 150 200
Heat flux (kW/m2)
18
Fragility Curves for On-grade Steel Tanks with % Full > 50%
Damage States
DS1 No damage to tank or I/O pipes
DS1
DS2 Damage to roof, minor loss of
contents, minor damage to
piping, but no elephant-foot
buckling

DS2 DS3 Elephant-foot buckling with

minor loss of content
DS4 Elephant-foot buckling with
DS3 major loss of content,
DS4 severe damage
DS5 Total failure, tank collapse

DS5

* Michael O'Rourke and Pak So, 2000

Structural Damage due to Earthquake 19
Structural Damage due to Earthquake

20
 Seismic Damage State vs Loss of Containment

Red color indicates 5%

plastic strain
Max = 189%
Elephant-foot buckling with major
loss of contentment,
severe damage
Damage = DS4
Rupture size = 11×4.5 m
Containment loss = 100%

Elephant-foot buckling
21
91
Thermal Radiation Contours due to Dike Fire at Tank 85 (10m/s Wind) -
93 95 97
78

Example 76

74

Thermal Radiation (kW/m2) contours

150

82 84 86 80
14
77
100 72

12 75

50
10

89
0 83 85 71 8 73

6
-50

4
-100
2

-150
-200 -150 -100 -50 0 50 100 150 200

22
 Thermal Radiation Contours for Tank Full Surface Fire (Example Case)
Envelop Max. Thermal Radiation from 5 Tanks

Random Variables:
- Wind Speed
- Wind Direction
- Tank Utilization

Event frequency = 1.2E-04 (OGP)

Thermal Radiation Contours at 20 seconds

23
 Individual Risk Contours for Tank Full Surface Fire

Random Variables:
- Wind Speed
- Wind Direction
- Tank Utilization

24
Individual Risk Contours without Domino Effects
Red color indicates 1.0E-04; Blue color: 1.0E-12
University and Residence: 1.0E-07 - 1.0E-06

25
Domino Effects

Domino system Heat Radiation

FARID KADRI, E. CHATELET, PATRICK LALLEMENT. THE ASSESSMENT OF RISK CAUSED BY FIRE AND
EXPLOSION
IN CHEMICAL PROCESS INDUSTRY: A DOMINO EFFECT-BASED STUDY. JOURNAL OF RISK ANALYSIS
26
 A Dike Fire that leads to Tank Full Surface Fires and Boil-Over- illustration

(1) Secondary containment area fire:

• Dike fire at Tank 104 (2)100
(2) Secondary event: 100
• Pool fire at Tank 100
(3)
103 (1)
104
• Boil-over at Tank 100 104
(3) Tertiary event:
• Pool fire at Tank 103
• Boil-over at Tank 103

101 102

Fire Impingement

27
 A Dike Fire that leads to Tank Full Surface Fires and Boil-Over

(1) Primary event:

• Dike fire at Tank 104 14

(2) Secondary event: (2)

12
• Pool fire at Tank 100 100 (1)

Thermal Radiation (kW/m2)

• Boil-over at Tank 100 (3)
103 104
10
(3) Secondary event:
• Pool fire at Tank 103 8
• Boil-over at Tank 103 101 102
6

28
Individual Risk Contours with Domino Effects
Red color indicates 1.0E-04; Blue color: 1.0E-12
University and Residence: 1.0E-06 - 1.0E-05

29
Conclusions
A comprehensive, holistic approach is required for determining risk in and around the
hydrocarbon storage tank terminal accounting for domino effects.
A proposed approach in this presentation closes the gap in the industry and combines the
knowledge of process safety, advanced structural analysis, and reliability together to
accurately and reliably estimate the individual, environmental and facility damage risks from
process and non-process hazards.
Analysis results show that domino effect may have a significant impact on the overall
risk calculation such that neglecting this effect can lead to overly low estimation in risk
studies.

30
THANK YOU!
Thank You! Questions??

Contact Details
Dr. Ali Sari, P.E.
Director
“Omega-risk”

Web site: www.omega-risk.com

E-mail: [email protected]
Potential Fire Scenarios at Tank Terminal

The most severe types are full surface fire and boil-over.
A full surface fire is where the tank roof has lost its buoyancy and some or all of the
surface of liquid in the tank is exposed and then involved in the fire
ARGYROPOULOS, C.D., CHRISTOLIS, M.N., NIVOLIANITOU Z., MARKATOS, N.C. “A HAZARDS ASSESSMENT METHODOLOGY FOR
LARGE LIQUID HYDROCARBON FUEL TANKS”. JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES. S.L. : ELSEVIER, 32
MARCH 2012. VOL. 25, 2, PP. 329-335.
Potential Fire Scenarios at Tank Terminal
Potential fire scenarios that can be developed in a tank accident include:

• Rim seal fire;

• Spill on roof fire;
• Full surface fire;
• Bund or dyke fire;
• Pontoon explosion;
• Boil-over.

Fire Types and Escalation

As shown in the diagram, relatively small fires initiated on the roof may lead to full
surface fire on a single tank. A full surface fire, if not extinguished, may lead to a
more severe fire scenarios such as boil-over on the same tank and/or full surface
fire on near-by tanks.

Shaluf, I.M., Abdullah, S.A., “Floating roof storage tank boilover”, Journal of Loss Prevention in the Process Industries, 2001.
33
Past Fire Hazard

Photos courtesy of Williams Fire & Hazard Control

Full Surface Fire

https://www.flickr.com/photos/dowgor/2969303992/in/photostream/ http://www.williamsfire.com/libraries/code_red/boilover_sleeping_giant.pdf

Secondary Containment Boil-Over

Pool Fire
34
Explosion (Blast)
Vapor cloud explosions can occur as a result of spillage of fuels or flammable
materials from aboveground storage tanks in tank farms.

Common causes that could unleash spills include:

• Overfilling
• Leaking from worn-out and corroded containment
• Loss of containment due to pipeline ruptures.
• Generation of combustible vapors that can be formed from the mixtures of
combustible liquids stored in storage tank.

35
Past Explosions

U.S. Chemical Safety and Hazard Investigation Board U.S. Chemical Safety and Hazard Investigation Board

Caribbean Petroleum Tank Terminal, October 23, 2009,

Bayamón, Puerto Rico

Buncefield Major Incident Investigation Board Buncefield Major Incident Investigation Board

Hertfordshire Oil Storage Terminal (Buncefield

Accident), December 11, 2005, Hertfordshire, England 36
Tornado Impact - Example
• Damaged Oilfield Tank on Briarwood after Tornado in Moore
Oklahoma May 2013

REFERENCE: GRAETTINGER, ANDREW J., ET AL. "TORNADO DAMAGE ASSESSMENT IN THE

AFTERMATH OF THE MAY 20TH 2013 MOORE OKLAHOMA TORNADO." (2014).
37
Seismic Hazard
• Seismic hazard is among the most natural threats to storage tanks. A seismic
event can lead to subsequent events (escalation) such as fire and explosion

Sari, et.al. (2007) Sari, et.al. (2007)

Tank farm fires after Kocaeli Earthquake Sloshing damage to upper shell and roof
(1999) in Turkey of tank during 1999 earthquake in
Turkey

38
Toxic Smoke

• Burning hydrocarbon fuel could generate

smoke and toxic gases.
• Consequence of these events on the
human health could range from
asphyxiation to burns.

• The smoke can be carried to long

distances away from the fire by the
wind and cause hazard to a wider
area.
• Absence of wind or presence of
heavy fog, can keep the smoke
close to the ground and therefore
hinder the firefighting operations or
endanger personnel.
Buncefield Accident

39
Toxic Smoke
• The hazard caused by smoke and toxic
gases is characterized by:
o plume dispersion rates
o height elevation
o the ground level concentration of
toxic pollutants.
International Business Times

o weather conditions Vasylkiv, Ukraine

• The smoke may contain toxic pollutants such as Sulphur dioxide (SO2), carbon
monoxide (CO), polyaromatic hydrocarbons (PAHs), and volatile compounds
(VOCs).
• Highly toxic Hydrogen Sulfide (H2S) could very quickly expose residential areas
to immediately life threatening conditions.
• Smoke containing Sulphur Dioxide (SO2) is potential health concern which could
be felt up to at least a couple of kilometers away.
40
Tolerability of Environmental Risk (Category Definitions) – Loss of Containment
Category Definitions
- Major airborne release with serious off-site effects
- Site shutdown
6 Catastrophic
- Serious contamination of groundwater or water course with extensive loss of
aquatic life
- Evacuation of local populace
- Temporary disabling and hospitalization
5 Major - Serious toxic effect on beneficial or protected species
- Widespread but not persistent damage to land
- Significant fish kill over 5-mile range
- Hospital treatment required
4 Severe - Public warning and off-site emergency plan invoked
- Hazardous substance releases into water course with 1/2-mile effect
- Severe and sustained nuisance, e.g. strong offensive odors or noise
disturbance
3 Significant
- Major breach of permitted emissions limits with possibility of prosecution
- Numerous public complaints
- Noticeable nuisance off-site, e.g. discernible odors
2 Noticeable - Minor breach of permitted emission limits, but no environmental harm
- One or two complaints from the public
- Nuisance on site only (no off-site effects)
1 Minor
- No outside complaint.
UK HSE, “Safety and environmental standards for fuel storage sites”, Process Safety Leadership Group, 2009. 41
Environment Agency for England and Wales, “Integrated Pollution Prevention and Control (IPPC) Environmental Assessment and Appraisal of BAT”, July 2003
Tolerability Criteria of Environmental Risk

Acceptable if reduced as low

Acceptable if Unacceptable if
Category Definition as is reasonably practical and
frequency less than frequency above
frequency between

6 Catastrophic 1.0E-06 per year 1.0E-04 to 1.0E-06 per year 1.0E-04 per year

5 Major 1.0E-06 per year 1.0E-04 to 1.0E-06 per year 1.0E-04 per year

4 Severe 1.0E-06 per year 1.0E-02 to 1.0E-06 per year 1.0E-02 per year

3 Significant 1.0E-04 per year 1.0E-01 to 1.0E-04 per year 1.0E-01 per year

2 Noticeable 1.0E-02 per year ~1.0E+01 to 1.0E-02 per year ~1.0E+01 per year

All shown as
1 Minor - -
acceptable

Broadly
Acceptable Region
Tolerable if ALARP Intolerable Region
Region
UK HSE, “Safety and environmental standards for fuel storage sites”, Process Safety Leadership Group, 2009. 42
Environment Agency for England and Wales, “Integrated Pollution Prevention and Control (IPPC) Environmental Assessment and Appraisal of BAT”, July 2003
Leak Frequency

Leak Frequencies Based on OGP Report No. 434-3

Leak Frequency
Type of Tank Type of Release
(per tank year)
Liquid spill outside tank 2.8E-03
Fixed/floating roof
Tank rupture 3.0E-06

43
Event Frequency
Event Frequencies Based on OGP Report No. 434-3
Fixed plus Internal
Floating Roof Tank (per Fixed Roof Tank (per
Event Floating Roof Tank (per
tank year) tank year)
tank year)

Full surface fire on roof 1.2E-04 - -

Internal explosion & full

- 9.0E-05 9.0E-05
surface fire

Large bund fire (full bund area) 6.0E-05 6.0E-05 6.0E-05

Event Frequencies Based on LASTFIRE

Event Frequency (per year)

Bund fire 1.62E-05

Full surface fire 4.21E-05

Boil-over Escalation probability 1.0??

44
Hazard Consequence Assessment Methods
Toxic
Fire Explosion
Smoke
Ref. NISTR 6546

Pool Fire: Solid Dispersion

TNO Multi
Plume, Point Models from
Energy
Source CCPS, DNV

Jet Fire model

TNT
from CCPS, CFD
Equivalency
DNV, API
Ref. NISTR 6546

Boil-Over Dilution Factor

model from UK BST vs Distance per
HSE OGP 434-7 OGP

CFD CFD

Rate of release (1)

Atmospheric condition (2) Plume Dispersion
Height of release AGL (3)
Distance from release (4) Model

Concentration
45
Tornado Fujita Scale Descriptions

46
Ref: https://en.wikipedia.org/wiki/Fujita_scale
Calculations Methodology for Loss of Containment
• In general, the total volume of liquid to be contained in a secondary
containment is considered to include three portions:
 Oil spilled from the tank;
 Rain water;
 Fire-fighting liquid (in a fire event).

• Probability of overfill due to Spilling:

F (OF Spill )  F ( Spill )  P(utilizatio n)  P( rain)

Frequency of Frequency Probability of Probability

overfill due to spill of spill tank utilization of rain event

47
General Risk Assessment Scheme Using Monte Carlo Simulations
Total individual risk, including domino effects, can be estimated using Monte Carlo simulations;
Random variables include the wind speed and direction, tank utilization, seismic hazard curves in the
case the primary event is an earthquake, tornado information (intensity, path, width) if the primary
event is a tornado;
Main steps in Monte Carlo simulations are illustrated in figure below:

Loop through all existing and new tanks at the site

1 2 3 4 5 6
Input data Random Primary event Identification of Escalation
variable consequence secondary/tertiary consequence
Risk
generation Assessment events assessment Calculation
(Combined
(Wind speed, (Flame Length, (Domino criteria, (Thermal Radiation, Thermal Radiation,
(Wind, Tank,
wind direction, Flame Tilt, consequence Toxic and Smoke, Individual Risk
Event
Frequency) tank utilization, Thermal vectors) Individual Risk) from each individual
PGA ) Radiation) tank)

48
Probability of Fatality using TNO Purple Book
Human Vulnerability – Toxic SO2 Exposure
1
5 min.
0.9 10 min.
30 min.
0.8 60 min.
120 min.
0.7 180 min.

0.6

Probility
0.5

0.4

0.3

0.2

0.1

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Concentration (ppm)

49
Individual Risk Contours for Toxic Smoke (CO and SO2)
Tank Full Surface Fire

50
Individual Risk Contours for Toxic Smoke (CO and SO2)
Secondary Containment Fire

51
Tornado Impact - Example
• Flipped Empty Oil Tank after Tornado in Edmonton in 1987

Ref: http://www.sherwoodparknews.com/2012/08/03/photographer-recalls-black-friday

52
 Terrorist Attacks
• Tank Failure, Baghdad, Iraq in May 2016 that was caused by terrorist attack.

At least 14 people were killed on Sunday in an attack by the militant group at a gas plant near Baghdad,
http://www.newindianexpress.com/world/2016/may/16/14-killed-in-IS-suicide-bombing-at-gas-plant-in-Iraq-904247.html
53
Risk Calculation Given No Escalation
𝑁

𝑃(𝐹𝐼𝐼) = 𝑃 𝐹𝐼 𝐴𝑖 × 𝑃 𝐴𝑖 × 𝑃(𝐼 𝑎𝑡 𝐴𝑖 )
Where, 𝑖

P(FII) = probability of an individual experiencing a fatal injury (FI);

P(FI|Ai) = probability of an FI occurring given an event (Ai) occurs;
P(Ai) = probability of an accidental event (Ai) occurring;
P(I at Ai) = probability that an individual is present when an accidental event Ai occurs;
N = number of events in fire, blast, seismic, etc.

The probability of individual present when an accidental event occurs will be estimated based on the working
schedules of the personnel on site, for example if an individual works 50 hours a week, this probability can be
estimated as 50/(7x24hrs) = 50/168 or 0.30. For the people outside the boundary of the terminals, this
probability will be assumed to 100%, i.e. 100% of the time an individual is present given an accidental event
occurs.

54
Risk Calculation Given Escalation
𝑁
𝑃(𝐹𝐼𝐼|𝐸) = 𝑖 𝑃 𝐼 𝐸𝑖 × 𝑃 𝐴𝑖 × 𝑃 𝐸𝑖 |𝐴𝑖 × 𝑃(𝐹𝐼|𝐸)
Where,
P(FII|E) = probability of an individual experiencing a fatal injury given an escalation;
P(I|Ei) = probability of an individual present given an escalation occurs;
P(Ai) = probability of an accidental event (Ai) occurring;
P(Ei|Ai) = probability of an escalation occurring given an event
P(FI|E) = probability of an FI occurring given an escalation occurs, i.e. human
vulnerability at certain escalation level;
N = number of events in fire, blast, seismic, etc.

55
Loss of Containment – Scenario 1

Area ≈ 50000 m2

56
Loss of Containment – Scenario 1
Thermal Radiation Contours (10 m/s Wind)
Thermal Radiation (kW/m2) contours
1200 14

1100
96 98
12
1000
78
91 93 95 97 76
900 74 10
79
800 88 90 82 84 86 80 77
72
75

700 8
87 81 83 85 89 71 73

600
6
500
10 m/s Wind
400 4

300
2
200

2200 2400 2600 2800 3000

PRELIMINARY RESULTS – DIKE FIRE
3200 3400 57