EFFECTS TRAINING: CONSEQUENCE MODELLING

02.12.2022
AGENDA
Day 1
• Introduction to consequence modelling
• Introduction to EFFECTS
• Key elements
• Functionality of user interface
• Theory of consequence models
• Release modelling
• Practical hands-on cases with EFFECTS

2
AGENDA
Day 2
• Theory of consequence models
• Pool evaporation modelling
• Fire modelling
• Dispersion modelling
• Practical hands-on cases with EFFECTS
Day 3
• Theory of consequence models
• Explosion modelling
• Damage models
• Receivers: vulnerable areas & congestion areas
• Combined models
• Practical hands-on cases with EFFECTS
3
GENERAL INFORMATION
In the shared folder you can find:

• Coloured books in PDF format

• Course presentations

• Background reader Part 1 to 3 → info about user interface, model physics & calculation models.

• Background Reader Part 4 → all exercises completely worked out (but don’t CHEAT)

• EFFECTS project files with all the exercises worked out

• RISKCURVES project files with all the exercises worked out

• An additional folder with some additional information

4
OBJECTIVES
At the end of the program, trainees will:
“be able to determine the consequences of the release of a hazardous material with EFFECTS”

Additionally, the trainees will:

• Understand the approach of performing consequence calculations
• Understand and distinguish several release modes and potential phenomena occurring (which model to use
and when)
• Be able to interpret and analyse the results of consequence calculations
• Be able to use the software tool EFFECTS for consequence modelling

5
 BACKGROUND: GEXCON, INSTRUCTORS AND TRAINEES

02.12.2022
ABOUT US
• World-leading company in the field of safety and risk management and advanced dispersion,
explosion and fire modelling.

• Our experience arises from detailed knowledge of explosion phenomena built up:
• through years of extensive research,
• carrying out assessments,
• performing accident investigations
• and conducting physical tests
OUR GOAL
• World-leading company in the field of safety and risk management and advanced dispersion,
explosion and fire modelling.
OUR PRODUCTS & SERVICES

SOFTWARE CONSULTING FIRE & EXPLOSION TRAINING

TESTING
ABOUT THE INSTRUCTOR
Sonia Ruiz Pérez – EFFECTS & RISKCURVES Product Manager
• Education
• Bachelor & Master in Chemical Engineering at IQS School of Engineering of Barcelona (Spain)
• Responsabilities at TNO & Gexcon (2014-Present)
• Research Scientist, Product Specialist and Product Manager
• Establishing the product vision, strategy and roadmap following discussions with internal and
external stakeholders as well as customers
• Expert on dispersion, explosion and fire modeling focusing in the development (and
implementation) of new consequence models in EFFECTS
• Responsible for verification & validation of new and existing EFFECTS models
• Providing training on consequence modelling, QRA and use of EFFECTS & RISKCURVES
• Performing QRAs and safety studies for the chemical industry using EFFECTS and RISKCURVES

10
ABOUT THE INSTRUCTOR
Hans Boot - Senior Research Scientist at Gexcon Netherlands
• Education
• Bachelor HTS Enschede (1984) Energy Technology
• Master University Twente (1988) Applied Thermodynamics
• Work experience
• Career at TNO, The Netherlands, since January 2019 Gexcon NL BV
• “Modelling thermodynamic processes”
• Research and Consultancy in
Industrial Energy conservation (1985 - 2003)
Industrial Safety (2003 - present)
• Responsible for contents EFFECTS and RISKCURVES software (incl. helpdesk)

11
ABOUT THE TRAINEES

?
Who are you ?
What is your background ?
What is your level of experience with consequence modelling and/or risk assessment ?
Have you ever worked with EFFECTS or other consequence modelling tools ?
Do you have any special requests?

?
12
 INTRODUCTION TO CONSEQUENCE MODELLING

02.12.2022
SAFETY ASSESSMENT
Why and how

Why do you want to perform a Safety Analysis ?

• From which perspective: Government / Company / Emergency response / Population
• A real awareness of danger or a regulatory ritual ?
• ALARA or ALARP: acceptable by who? Government? The people around it? → That’s why QRAs are
required.
“as low as reasonably practicable”
How: derive quantitative values, but: which units to use as criteria ?
• LEL, IDLH concentrations, # of victims, % lethality, individual risk, societal risk
• Deterministic approach: how far reaches effect
• Probabilistic approach: what is risk: involves chance

14
SAFETY ASSESSMENT
Communication

Communication about Safety / Risk is very sensitive!

• Even talking about potential lethality, damage or injuries scares people!
• Awareness of risks is often low
• The concept of risks is barely understood
• Chance of dying by smoking vs. chance of dying by a nuclear accident?

How to balance safety against efficiency / competitiveness?

• Translate Risk into Costs: product loss, production capacity loss, installation/ equipment, reputation
damage

15
IT STARTS WITH LOSS OF CONTAINMENT
WHAT IS CONSEQUENCE MODELLING

Predict the consequences of an accidental release of a hazardous material to calculate the

consequences that a fire, explosion or dispersion scenario have on surrounding population
and structures.

So that we can understand the consequences of an accidental release, to design and

operate safe facilities (i.e., implementing prevention and mitigation measures).
 QRA
CONSEQUENCE VS RISK Consequence
analysis
Define potential
Consequence analysis: accident scenarios

• To calculate the consequences

• What could happen? Evaluate event
consequences
Estimate potential
accident frequencies

• If it happens, how dangerous is it?

• For ONE weather class, wind speed… Estimate the impact
of events
• Potential part of QRA (Quantitative Risk Assessment)

Estimate the risk

Risk modelling:
• Risk = Likelihood ∙ Consequence
Evaluate the risk

• Uses failure frequencies & weather statistics Risk

modelling
Identify and prioritize
potential risk
reducing measures

18
CONSEQUENCE AND RISK MODELLING MADE EASY

EFFECTS RISKCURVES
Advanced yet user-friendly software to analyse Advanced software for performing QRAs for
the effects and consequences of the accidental storing and transporting dangerous goods in
release of dangerous chemicals. process and (petro)chemical facilities.

Calculates heat radiation from fires , over- It presents location-specific individual risk
pressures from explosions , toxic contours, fN curves and societal risk maps, and risk
concentrations and dose from dispersion . ranking reports.
LITERATURE SOURCES
• EFFECTS & RISKCURVES based on “coloured Guidelines for the use of RISKCURVES
books”. Purple Book (NL: Handleiding Risicoberekening BEVI)
(CPR-18)
HAZID, HAZOP, FMEA Identification of
Chance scenario
accident scenario’s

Determining physical
Red Book
Yellow Book (CPR-14) Chance effect (CPR-12)
effects
Contains statistics
& failure frequencies
Determining the
Green Book (CPR-16) damage
Chance damage

Consequence of risk Chance of risk

• Other sources: Consequence Chance

• CCPS Guideline books Quantification of risk

• HSE publications
Evaluate risk
• Local regulations
20
Risk Reducing proposals
 INTRODUCTION TO EFFECTS

02.12.2022
HISTORY
1980’s
1990’s
2000’s
2010’s

2020

2022

22
WHAT CAN EFFECTS DO
EFFECTS is used to “predict the consequences of an accidental release of a hazardous material” and is
aimed at flammable, explosive and/or toxic materials.

EFFECTS is intended for:

• Calculation of leak rate / evaporation rate (= outflow determination)
• Calculation of damage to people and constructions (due to heat radiation, overpressure and/or toxic
exposure)
• Calculation of consequence effects, which is typically part of a “Safety Assessment” (Preparation phase).

EFFECTS is not intended for:

• Radiologic or Biological hazards
• Long term exposure effects (e.g. pollution)
• Assessment of the “repression phase” (e.g. emergency response)
23
EFFECTS VS RISKCURVES

Physical Effects Consequences Scenario’s & frequencies

EFFECTS - consequence analysis RISKCURVES – quantitative risk analysis

• Single models, per phenomenon • Multiple scenarios, per installation
• How far reaches a specific EFFECT • What is cumulated RISK (chance * consequence)
• Usage: process (safety-)analysis, validating Includes “chance” of accident, weather statistics,
legislative requirements, emergency population-distribution: more complex input!
response, Preparation QRA scenarios! • Usage: legislation, urban development, risk
reduction (external and internal safety)
RISKCURVES calculation core uses EFFECTS consequence models 24
TYPICAL USAGE
Including translation
to damage
• Heat radiation levels (kW/m2) • First/second/lethal degree burns
• Overpressure levels (mbar) • Fraction of mortality indoors/outdoors vs distance
• Toxic levels (ppm) • Fraction of mortality at study height
Values
Graphs
Maps

25
THE MODELS

• One user interface with around 40 different physical models

• Basic background literature:
• Yellow book → for the calculation of physical effects
• Green book → for the translation from consequence effects to damage (in number
of fatalities)

• Models remain an approximation of reality.

26
THE MODELS FOR CONSEQUENCE ANALYSIS
Requires some background knowledge:
• Which effects/phenomena models to choose?
• What is the validity for this application? This is why you’re here
• What parameter values should I choose?

Typical phenomena:
• Fire: jet fire, pool fire, flash fire, local cloud fire, BLEVE fireball, warehouse fire
• Explosions: overpressure and pressure impulse
• Toxic: Dispersion poisonous cloud → dose → lethality
• Rupture of vessels: Fragmentation, projectile throwing range, glass damage
These phenomena are usually location and time dependent → results need to be presented in graphs
27
 WHAT’S NEW IN V12
• Extended dispersion model with automatic dispersion behaviour detection. Allows the simulation of all
dispersion behaviours including: dense, neutral and lighter than air.
• Improved high-speed jet releases, which is highly relevant for H2 releases (commonly stored and
transported at high pressures – up to 900 bar)
• Improved modelling to account for ground effects and plume reflection
• Fire models have been adjusted to account for hydrogen’s unique properties (e.g., max 70 kW/m2)
• Modernized user interface making it more user-friendly
• Import footprints from other consequence modelling tools into RISKCURVES to calculate accumulated risk
• New cloud licensing

28
 THE EFFECTS USER INTERFACE

02.12.2022
HOW DOES EFFECTS WORK
• A consequence model requires a specific set of inputs:
• Type of chemical (DIPPR substance database), dimensions, operating pressures, environment
conditions, etc.
• Typical output will contain:
• Values (max. concentrations, heat radiation, lethality, etc.)
• Graphs (value vs distance, value vs time)
• Maps (contours, value distributions), shown on top of background topographic map
• Models can be linked creating a chain : output outflow = input dispersion
Flow rate,
Dimensions, Exit conditions
Pressures, Outflow Concentration contours,
Temperatures Lethality footprint,
Dispersion Concentration vs distance,
Concentration vs time
Meteo data
30
EFFECTS V12 USER INTERFACE
Menu bar Toolbar Result screen tabs (values/graphs/maps)

Model selection /
Project tree

Model input panel

Result
display area

Command button INPUTS RESULTS

panel
31
LET’S START USING EFFECTS
• Start EFFECTS
• Select a model (from menu or “auto-hiding” side panel or right-click project tree)
• Select the BLEVE fireball in “Fire” models group
• Click “Defaults” and click “Calculate”:
• Yellow → Warning (e.g. extrapolation outside range)
• Red → Error (e.g. input field empty)
• No light → OK
• Evaluate Results / Profiles / Contours and Report
• Other graphs: Use “graph selection” combo-box

32
INPUTS & RESULTS SCREEN
• Each model has its own list of input & result parameters
• Input → Edit boxes & Combo boxes (Unit conversion, Colour highlighting)
• Choices made in combo boxes might modify list of input
(some models have “sub-models”: instantaneous or continuous)
• Tabs: Results, Graphs, Contours, Reports
• F1: help

33
 MODEL COMPARISON
• We often apply “parameter variation” evaluations:
• What if we had a bigger vessel?
• What if we had stronger wind?
• What if we had summer temperatures?
• The model tree allows to simply “copy/paste” a model Ctrl-C + Ctrl-V
• We copy the model to the same location … or … we copy the location (equipment point) & change
coordinates
• We modify input and adjust description of the node (F2)
• After calculation (Italics or Out of date) : alternative results
• The selected node of tree determines results shown
• To illustrate differences/multiple results: Add a “Comparison set”

34
 GRAPH FUNCTIONALITY
• Profile selection • Unit conversion

• Crosshair & Ruler

• Profile Expert: all available graphs

• Zoom / scroll

• Properties (to edit graph), Copy, Freeze, Crosshair

35
MAP FUNCTIONALITY
• Contours illustrate scale of event

• Commonly applied on top of background map (tile-server/file based)

• Toolbar: Show grid, Release point, Crosshair, Ruler, Lock scale vs Full extent

• Zoom / scroll

• Legend options: (de-)selection, grid colour coding,

& colours for contours, export and copy layer, freeze layer

• Map: scale bar, north arrow, unit conversion

• Export: per layer or full image

• Show surrounding contours

36
MODEL LINKING
To create a “chain of models”

• Models can be linked individually → Output of model 1 = Input model 2

• Model tree “Copy & Paste” action → Copy model 1 and paste on top
• Alternatively use <right mouse> → Link to / Unlink

• Linked parameters will be illustrated in blue, and can be relinked <right mouse>
• Overwriting a value will UNLINK: black parameter caption

Also very useful to synchronize models: force to use same inputs!

37
USEFUL TOOLS
Edit → Options (all defaults):
• To indicate default expert parameters
• Environment
• Dispersion
• Display units
• Location of chemical databases
• To change language
Tools:
• Mass & volume calculator: converter Mass ↔ Volume
• Mortality/Probit calculator: concentration + time → dose
• Probit converter: unit converter

38
PROJECT WORKFLOW
Use of the tree: work from top to bottom

• Presentation settings → projection system, contour levels, colours to use

(What coordinate system to use? Which level for contours to show?)
• Backgrounds → to load tile server or background map
• Receivers (optional) → vulnerable area for damage, congestion area for VCE
• Equipment → to pinpoint location (coordinates) of equipment on background map
• Add models at equipment → copy / paste from other
• Make comparison sets → evaluation, comparison of results

39
PRESENTATION SETTINGS
• Define (coordinate) projection system
• Relative or official system?
• Country dependent:
• Norway → NGO 1948
• Belgian → Lambert
• Australia →
• Which typical contour levels to show?
• Heat radiation
• Overpressure
• Concentration
• Lethality
• What colours to use for graphs and “footprint” grid distributions
• Changes don’t require recalculation! (just visualisation)
40
BACKGROUNDS IN EFFECTS AND RISKCURVES

• EFFECTS supports multiple backgrounds, which are stored inside the project file to allow easy exchange of
projects between PC’s.

• Backgrounds can be files (images / AUTOCAD drawings),

or directly taken from the internet

• Backgrounds can be at absolute positions to facilitate integration of results in GIS systems or Google Earth.

• It is still possible to work with relative

positions around (0,0) if results don’t need to be exported.

41
BACKGROUNDS: TILE SERVER
• What → a set of images (tiles) retrieved from an internet connected server
• Why → always a correct scale and location
• Why not → internet connection dependent / sluggish behaviour / zoom-zoom-zoom
• But → a specific viewport can be saved including scale + projection
• And then → multiple saved viewports (files) can be combined

• How → <Add tile server>, zoom in and lock scale ( )

• Just a few choices yet (OpenStreetMap)?
• ESRI / Google / MapBOX are HQ licensed options, but ……
42
BACKGROUNDS: TILE SERVER
• We are not allowed to include specific servers in a “commercial” tool
• Some useful “satellite view” tile servers are “open” or allow trying out
• Option 1: ESRI World Imagery (see “additional tile servers.txt” on USB)
http://services.arcgisonline.com/arcgis/rest/services/World_Imagery/MapServer/tile/{level}/{row}/{col}.png (PNG
format)
• Option 2: Google Satellite / Hybrid / Maps
http://mt1.google.com/vt?lyrs=s&x={col}&y={row}&z={level}
http://mt1.google.com/vt?lyrs=y&x={col}&y={row}&z={level} (PNG format)
• Copy / Paste URL addresses
• Enter Name / Type / Preset
• Press Save . . .

43
EXERCISE: ADD GOOGLE HYBRID

44
EXERCISE: ADD GOOGLE HYBRID

1. Add background map → Right-click mouse and browse for “add tile server background map”)

2. Select Google Hybrid → Press OK

45
BACKGROUNDS: FILE BASED MAPS
• Projection system: contains location of (0,0) coordinate + rotation to North.
The projection system is the first thing to decide when using maps
• Vector maps include a scale (and may include coordinate system):
Shape files SHP + DBF (+ PRJ)
AUTOCAD DXF (+ PRJ)
• Pixel oriented maps: JPG, PNG, TIF, GIF and BMP do not include coordinates.
A “GeoTIF” already contains projection system.
• How many pixels is how many meters? Absolute location of a pixel: stored in ascii “world file”
JGW/PGW/TFW
“Geo-referencing” of pixel-oriented maps (*.jgw/pgw/tfw, *.prj)
• Can be used with Google Earth “screen capture” !
• A tile server loads maps from internet, but tiles “field of view” can be saved as PNG (+ PGW/PRJ) or JPG
(+JGW/PRJ)
46
WHY DO G.I.S SYSTEMS NEED A PROJECTION SYSTEM?
Projection system → Which coordinate system to apply
• Projection system describes HOW to project the earth on a 2D surface:
• Conversion Lat/Long degrees to (x,y) in Map units (any unit possible, often [m])
• There are hundreds of different projection systems
• Default: “Pseudo Mercator”: Unit meters relative to Lat/Long 0,0 = Greenwich / Equator
• Belgium: “Belgian Lambert”, Singapore: SVY21, Netherlands: Amersfoort RD, Norway: NGO 1948,
Australia: GDA94

47
BACKGROUNDS
Working with geographic locations
Absolute system Relative system
Description Uses “real” locations as used in GIS systems. Uses locations relative to the origin (0,0).
• Select appropriate projection system. • Select the “Pseudo Mercator” projection system.
• Select background: tile server or geo-referenced image • Select background: Geo-reference background(s) taken
How? file(s). from Google Maps and set one point to (0,0).
• Input equipment locations: by locating them on map or • Input equipment locations: by locating them on the map.
use map coordinates from used projection system.
• Results on the “map tab” are correct and can be used in • Results on the “map tab” are correct and can be used in
reports. reports.
Result
• Exported contours / grids can be imported in GIS • Exported contours / grids cannot be imported in GIS
systems/Google Earth. systems/Google Earth.
• Tile server
Examples • AUTOCAD drawing (vector based)
• Geo-referenced image file (pixel based)

48
EXERCISE: GEO-REFERENCE A JPG
BACKGROUND MAP

49
EXERCISE: GEO-REFERENCE A JPG BACKGROUND MAP
Practice geo-referencing with “Barcelona GoogleEarth.jpg” (in “EFFECTS 12 Training Example files” folder in
USB memory stick).

1. Add background map (right-click mouse and browse for map)

2. Select default coordinate system WGS 84 (this file has no rotation)
3. Select <Geo-reference by length>
• Zoom-in (mouse wheel), select point at “Lat Long” pin
• Move to bottom left (right-click mouse)
• Select begin/end length indicator (total 3 red points)
• Enter Lat 41.351485° Long 2.159574° (≠ Map units)
• Enter Length as displayed: 900 m
4. Check geo-referenced map with OSM tile server

50
EXERCISE: EXPORT A GEO-REFERENCED
PIXEL BASED MAP FROM A TILE SERVER

51
EXERCISE: EXPORT A GEO-REFERENCED PIXEL BASED MAP
FROM A TILE SERVER
Tile server requires internet → might be slow. Solution: export geo-referenced pixel-based-map from tile server.

1. Add tile server (Right-click in “Backgrounds”)

2. Select tile server (under “Preset”)

3. Export tile server image as geo-referenced pixel based map

• Zoom-in in area of interest → Right-click on map → Export → Export to file → Save as pixel based file (e.g.
png)
4. Import pixel based map (e.g. png) as background map
• Add background map (Right-click in “Backgrounds”)
• Select the previously exported pixel based file (which already contains a coordinate system) → Click OK
More info in user manual

52
EXERCISE: EXPORT YOUR CONTOURS TO
GOOGLE EARTH

53
EXERCISE: EXPORT YOUR CONTOURS TO GOOGLE EARTH

• Combine tile server + geo-referenced file → Does it fit?

• Select an equipment location at sphere on the site
• Activate equipment node
• Zoom-in (wheel) → select point (right-click) → “Set release point”
• Calculate a BLEVE fireball at chosen location
• Evaluate heat radiation / damage circles
• Select legend layer “Contours”
• Right-click → “Export to file”
• Save as KML (potentially as .SHP)
• If Google Earth installed: double-click KML saved file

54
CHEMICAL DATABASE
What is it? (1)
• The chemical database is an extensive database of rigorously evaluated pure component data on industrially
important chemical compounds.

• The chemical database can be found in both EFFECTS and RISKCURVES.

• Chemicals play a major role as they determine consequences.

• But then we need substance specific “properties”: LFL, boiling point, vapour pressure etc.

55
CHEMICAL DATABASE
What is it? (2)
• Setup database locations → Chemical database manager (Edit-Options)
• Multiple files possible / can be combined
• Std EFFECTS database (Chemicals.tci)
• User defined database
• Corporate database (on server?)
• Client database
• Access database editor by toolbar button or behind Chemical
• Default sources: DIPPR 2015, DIPPR 2010, YAWS, Sample mixtures
• DIPPR: Licensed, >1500 chemicals, non ideal gas behaviour (compressibility)
• Any addition/modification will be stored as DIPPR EDITED record (in user defined file)

56
CHEMICAL DATABASE
Editor (1)

Filter by...
• Name (synonyms, full CAS number,…)
• Flammable or toxic
• Source (DIPPR, YAWS…)
• Properties (liquid density, molecular weight,…)

57
CHEMICAL DATABASE
Editor (2)
Adding parameters:
• Copy chemical to editable database
• Added chemical preceded by orange bullet

• All properties in the new chemical preceded by orange bullet

• Rename (traceability)

• Add property to the new generated chemical

(e.g. threshold, toxic Probits, LFL, …)

58
CHEMICAL DATABASE
Editor (3)
Modifying existing parameters:
• Copy chemical to editable database
• Modify added chemical (e.g. local toxic Probit instead of Purple Book Probit)

Important for traceability: use a descriptive name for the new chemical: e.g. “Ammonia French Probits”

59
CHEMICAL DATABASE
Editor (4)

Chemical is defined by…

• Constant properties (e.g. LFL, UFL, Tb, Molar weight)

• Temperature dependent properties (e.g. PvSAT vs T)

• Which have a valid temperature range

• Threshold properties

(e.g. AEGL, ERPG, IDLH VRW, AGW, LBW)

• Toxic properties (e.g. Probits A, B, N)

Open ICSC button (Internet required)
links to important source !

60
CHEMICAL DATABASE
Mixtures
• Some sample mixtures supplied
• But sample composition may deviate from local composition

• Calculated properties mixture based on mixing rules

• E.g. Ideal mixing, Le Chatelier, Kendall-Monroe, etc.

• Limitations:

• No binary interaction, No chemical reactions, No absorption effects, No solubility

• We assume constant composition (also within pool evaporation)
• The chemical for dispersion has same composition as liquid in pool (when linked)
• Based on quadratic EOS including compressibility based on second virial coefficient
(super critical conditions are unreliable)
61
CHEMICAL DATABASE
Creating your own mixture

1. Search for component (filter name, source)

2. Right-click → Select <Add to mixture> or <Drag> (mouse) to mixture panel
3. Define Parts (can be mole based or mass based)
4. When finished, press <Apply> and give name
5. It will be stored in User defined database
6. A “Writable” user database must exist!
(Menu: Edit → Options → Chemical database manager)

62
EFFECTS PROJECT FILES

• We work with project files (.effx = zipped XML-format file)

• One project file can contain multiple models

• Each model can contain multiple calculations (sessions): different parameter sets

• A project file stores input, maps, chemicals and calculation results

• Clean-up your project from unintended model calculations (Use <delete> on selected model in tree)

• Use descriptive names for your model names (Use <F2> or node click to rename in tree)

63
EFFECTS PROJECT FILES

• Each model has a unique model version and title

• The program itself contains a version reference info for every model (that refers to the technical
documentation)

• New EFFECTS version: new model version: possibly different answers.

• Check version information for changes! Menu: Help → Version information

64
 THEORY OF CONSEQUENCE MODELS

02.12.2022
AVAILABLE MODELS IN EFFECTS
1. Release models → Outflow and Evaporation

2. Dispersion models → Plume rise, Neutral and Heavy Gas: Toxic dose, flammable cloud &
concentration

3. Fire models → Pool fire, Jet Fire, Flash Fire, Gas and BLEVE Fireball, Combustion products

4. Explosion models → TNT (solids), Multi Energy (VCE), Vessel Rupture, Gas and BLEVE Blast

5. Damage models → Translation from consequence to lethality

6. Combined models → Predefined chain of models, autopilot mode:

→ Release → Evaporation → Dispersion → Toxic cloud → Lethality
66
AVAILABLE MODELS IN EFFECTS

67
WHY SO MANY MODELS ?
• Different situations: instantaneous (catastrophic failure) vs continuous (leak)
• Different states: Gas, Liquid, Pressurised Liquefied Gas (PLG)
• Different chemical hazards: flammable or toxic material
• Different phenomena: fire, explosion, toxic exposure
• Sometimes a chain of models is required and/or multiple phenomena are possible:
• Gasoline leak (flammable): Outflow → Pool evaporation → Dispersion → Explosion of flammable cloud
• Acrolein leak (both flammable and toxic):
• Direct ignition: Outflow calculation → Pool fire
• Delayed ignition: Outflow → Pool evaporation → Dispersion → Toxic exposure / Explosion of
flammable cloud
• You should have a feeling of what to expect for specific situations (although combined models will detect this)
68
WHAT PHENOMENA TO EXPECT ?
• Instantaneous/continuous release of liquid
• Distinguish “direct” and “delayed ignition”

69
WHAT PHENOMENA TO EXPECT ?
• Instantaneous/continuous release of gas
• Distinguish “direct” and “delayed ignition”

70
WHAT PHENOMENA TO EXPECT ?
• Instantaneous/continuous release of pressurized liquefied gas
• Distinguish “direct” and “delayed ignition”

71
 OUTFLOW MODELS IN EFFECTS

02.12.2022
INTRODUCTION
Important things to consider:
• Phase of chemical → Models for liquid, gas, pressurised liquefied gas

• Location leak:
• In vessel or pipe
• If length pipe > 1 km → use “long pipeline” model)
• Height of the leak: in liquid or in gas phase?

• Type of release: catastrophic failure, full-bore rupture, small leak

• Release driven by (hydrostatic) pressure or pump?

73
 TYPES OF OUTFLOW MODELS
Release models

Gas release Liquid release

Liquefied gas release

Gas release Gas release Gas release long Liquid release

Gas release Liquid release Liquid cascade
instantaneous 10 minutes pipeline 10 minutes

Outflow models typically

2-Phase 2-Phase bottom 2-Phase long 2-Phase 10 2-Phase include vessel dynamics:
vapour venting discharge pipeline minutes Instantaneous time dependent flowrate.
Spray
Release Rainout mass rate
Pool
Evaporation 74
GAS RELEASE
• Gas release: hole in vessel / pipe → A, B, C

• Gas release in 10 min: entire contents are released in

10 min (calculates corresponding hole size)

• Gas release from long pipeline:

for pipes >1km. Non-stationary model that calculates
effects of “pressure wave” → D, E

• Gas instantaneous release: predicts gas expansion and cooling behaviour

(relevant for delayed ignition situations) → F

75
LIQUEFIED GAS (TWO-PHASE) RELEASE
• Liquefied gas vapour venting: for top venting releases. Includes liquid “bubbling effects” → B, D, F

• Liquefied gas bottom discharge: hole in bottom of vessel/pipe. → A, C, E

• Liquefied gas instantaneous release: for catastrophic rupture. Instantaneous flashing release. Calculates rain-
out + airborne mass → J

• Liquefied gas spray release: to be used after TPDIS model to

calculate rain-out, liquid fraction and airborne mass

• Liquefied gas from long pipeline: for non-stationary

releases. Calculates effects of “pressure wave” inside
pipe → G, H, I

• Liquefied gas release in 10 min: 10 min (calculates hole size)

76
LIQUID RELEASE
• From vessel or pipe: hole in vessel / pipe. Straightforward Bernoulli

• In 10 min: entire contents are released in 10 minutes (calculates corresponding hole size)

77
LIQUID RELEASE – PB APPROXIMATION
Most release models provide a “time dependent source rate” q (kg/s) vs t (s) but …
What is the correct “Representative Source Rate” to be used use as dispersion rate? → PURPLE BOOK APPROXIMATION
• The PB approximation is a translation of time dependent flowrate into semi-continuous flowrate and duration
• Method applicable for release models and pool evaporation model
• Methodology (included in EFFECTS) 1st step representative rate

• Calculate the total amount of mass released in first 30 min Time at which representative
T/P/etc. is taken
• Divide in 5 blocks / steps
• Each block has 20% of total released mass 2nd step representative rate

• Cut-off at maximum 1800 s

• Representative rate
Area = 20% of released mass
• Flammable → 1st step
1/5 of representative duration
• Toxic → 2nd step (in case of use 2nd step)

78
LIQUID RELEASE – INPUT PARAMETERS
• Driving force for outflow models → Pressure
• For liquids: liquid height (hydrostatic pressure)
• For gases: overpressure or vapour pressure
• For liquefied gases: overpressure above the liquid (typically the vapour pressure or actual pressure)

• Typical input parameters:

• Type of vessel: determines the liquid height or contents (volume/mass)

79
LIQUID RELEASE – INPUT PARAMETERS
• Typical input parameters:
• Type of release: through hole in vessel, through (hole in) pipe, fixed feed

• Initial pressure: vessel dynamics are included

• Hole rounding (discharge coefficient Cd): ratio Doutflow / Dhole
• Guillotine fracture pipe → rounded edges, Cd = 1 Dhole Doutflow Dhole Doutflow
• Leak in vessel/pipeline → sharp edges, Cd = 0.62
• Roughness pipe: influences friction loss in pipe (click <F1> for typical values from YB)

80
LIQUID CASCADE NEW!

• Tank overfill incident used to be treated as simple liquid spill

• Model based on HSE publication that follows investigation of Buncefield incident

• Vaporization occurs from the small liquid droplets formed in the cascade, forming a highly flammable cloud

• Extended to allow:
• Calculation of concentration at foot of the cascade for all chemicals
• Can be linked to dense gas dispersion, explosion and pool fire models
• Calculates evaporation from cascade and generated liquid spill

81
 EXERCISES OUTFLOW MODELS:
EXAMPLES FOR RELEASES OF
LIQUID / GAS / LIQUEFIED GAS (EXERCISE 1 TO 5)

82
LESSONS FROM OUTFLOW EXAMPLES
• Liquid release:
• Driving force → liquid height (=hydrostatic pressure) → vessel shape
• Liquid height (vertical vessel) > Liquid height (horizontal vessel/sphere)
• Liquid height (vertical vessel) decreases faster than Liquid height (horizontal vessel/sphere)
• If pipe → ↑Pipe friction → Flowrate↓
• Gas release:
• Driving force → Overpressure / vapour pressure
• Initial pressure (overpressure/vapour pressure) → gas flows out → P and T drop → Flowrate ↓
• If pipe → ↑Pipe friction → Flowrate↓
• Two-phase release:
• Driving force → Overpressure above the liquid (vapor pressure)
• If pipe → liquefied gas expands → vapour creation → less space for the liquid → Flowrate↓

83
LESSONS FROM LIQUID CASCADE EXAMPLES
• Liquid cascade:
• Calculates the evaporation rate from the liquid cascade
• “Vaporization rate from liquid cascade”
• “Density after mixing with air due to vaporization from cascade”

• Calculates the evaporation rate from the generated liquid spill

• “Purple Book representative pool evaporation rate”
• “Density after mixing with air due to pool evaporation”

• Cumulates all evaporation rates which are used as source rate for the dispersion
• Evaporation from cascade + splash products + pool

84
 EVAPORATION MODELS IN EFFECTS

02.12.2022
TYPES OF EVAPORATION MODELS
Flash evaporation Evaporation of liquid droplets in Pool evaporation (of liquids or
aerosol rain-out fraction)

EFFECTS model Liquefied Gas Instantaneous Liquefied Gas Spray Release Pool Evaporation model
Release (AMINAL) model model
Suitable for Evaporation of an instantaneous Evaporation of a release of Evaporation from a liquid pool
release of pressurized liquefied pressurized liquefied gas from a spreading on land or water
gas (flashing liquids) vessel or pipe with Tb < Ta, or
champagne release.
Calculates Adiabatic vapour flash fraction, Adiabatic vapour flash fraction, Evaporation rate and
liquid fraction airborne (droplets) liquid fraction airborne (droplets) evaporation duration
and rain-out and rain-out

86
LIQUEFIED GAS INSTANTANEOUS RELEASE MODEL
• Aim: How much liquid ends up in the pool? And as an aerosol (liquid droplets) in the cloud?
• Based on rule of thumb → AMINAL model for instantaneous releases
• Adiabatic flash (χ) → amount of liquid mass that has been instantaneously evaporated upon expansion to Patm
• Spray fraction (δ) → fraction of inventory dragged into the cloud as liquid droplets
• Airborne mass (ABM) → Total mass in the cloud (both liquid and vapour) after instantaneous release = χ + δ
Adiabatic flash fraction (χ) Spray fraction (δ) Airborne mass (ABM) = χ + δ Comments
χ < 0.10 δ=χ ABM = 2 * χ
0.10 ≤ χ < 0.36 δ = (0.54*χ – 0.028)/0.26 ABM = (0.8*χ – 0.028)/0.26 2* χ < ABM < 100%
χ ≥ 0.36 δ = 1- χ ABM = 100% Nothing rains-out

• Inst. liquid → Direct ignition → BLEVE

→ Delayed ignition → Dispersion
• Also provided by the model: mass in pool, representative density of aerosol, radius of expanded cloud
87
LIQUEFIED GAS SPRAY RELEASE MODEL
Functionality
• For releases of pressurized liquefied gases
• ALWAYS used after the two-phase bottom discharge model (TPDIS)
• When jetted into air, evaporation and expansion creates vapour + small liquid droplets
• Part of the droplets are dragged along with vapour (=spray) → into airborne mass
• Part of the material rains-out → possible pool-formation
• Expansion of liquid will create expanded jet

Doutflow Dexpansion

• Density of expanded cloud influenced by liquid fraction (droplets)

• The density of expanded cloud tells whether the dispersion phenomena is heavy
• Main answers: rain-out mass, airborne mass, expanded diameter, density aerosol (heavy & cold)
88
LIQUEFIED GAS SPRAY RELEASE MODEL
3 models for Spray Release
EFFECTS has 3 different models for Spray Release modelling:
• AMINAL rule
• Based on rule of thumb (no physical phenomena)
• Developed for Flemish governmental regulations

• Yellow Book (recommended model)

• Based on physical phenomena and average particle size

• Statistical Spray (recommended model for CO2 releases)

• Based on physics
• Developed for CO2 modelling
• Accounts for droplet breakup and randomized droplet size distribution
89
POOL EVAPORATION MODEL
Evaporation highly influenced by:
• Liquid vapour pressure → ↑T, ↑Pvap, ↑evaporation rate
• Wind speed → ↑wind speed, ↑evaporation rate
• Pool size (bund vs no bund)
• If bund → pool size limited to area of bund
• If no bund → pool has no restriction on size
• Type of release
• Instantaneous → content is released in short time
• Semi-continuous → size of the pool determined by flow rate (slower growth)
• Heat transfer rates

90
POOL EVAPORATION MODEL
Evaporation by heat transfer from:
• Internal energy → inlet flow at higher temperature. Can be warmer than evaporating pool
• Convection → Air and water vapour in air
• Radiation → Solar heat radiation
• Conduction → Subsoil type: land vs water.
• Pool on water:
• Constant water circulation → ↑∆T (Twater - Tpool) → ↑heat transfer
• No limitation on pool size → Pool thickness small & ↑contact area pool → ↑heat transfer

91
POOL EVAPORATION MODEL
Boiling vs non-boiling pools
Boiling pools (A, C).
E.g. LNG, pool at -161°C (Tb)
Evaporation highly influenced by…
• Heat transfer between subsoil and pool
• Subsoil type (land vs water)

Non-boiling pools (B, D).

E.g. Acrolein, pool at Tstorage (< Tb = 52°C)
Evaporation influenced by…
• Wind speed
• Solar heat radiation

92
POOL EVAPORATION MODEL
Other input parameters (1)

• Maximum pool surface area

Only used when the pool is bunded.

Pool size estimation based on real world:

• Use (satellite) background
• Measure area (toolbar “ruler” icon)
• Draw area for realistic pool size: area in m2 shown

93
POOL EVAPORATION MODEL
Other input parameters (2)
• Subsoil roughness
The pool will spread until it reaches a minimum thickness. When this minimum thickness (= average
roughness) is reached, the pool size will start decreasing in size.
This minimum thickness is related to the surface roughness.
Subsoil Average roughness
Flat sandy soil, concrete, tiles, plant-yard 0.005 m
Relatively flat sandy soil, gravel 0.010 m
Rough sandy soil, arable land, meadows 0.020 m
Very rough overgrown sandy soil with holes 0.025 m

• Type of subsoil
Material in which the pool spreads on. Determines the heat transfer behaviour.
94
POOL EVAPORATION MODEL
Output parameters

Density after mixing with air (=density of the cloud after the vapour from pool evaporation has mixed with air)
• What is now density mixture ? Is it ≈ to density air ?
• As of v12, you no longer need to decide what is the expected dispersion behaviour of the cloud, but this
result parameter is useful to understand what to expect.

Amount of air mixed with vapour cloud is calculated as:

Wind speed [m/s] * Width pool [m] * 0.5 m * Density air (kg/m3)
The density of the mixture after mixing with air is determined by this mass flow of air mixed with the cloud

Representative evaporation rate

Source rate of the dispersion → calculated with Purple Book 5 step approach

95
POOL EVAPORATION MODEL
Limitations

Behaviour NOT INCLUDED in the EFFECTS pool evaporation model

• Substance may dissolve (e.g. methanol) or react (e.g. ammonia)

• Substance may float or sink

• Substance may create an ice layer while evaporating (e.g. LNG on very still water)

• Evaporation rate of a mixture is calculated assuming an average mixture Tboiling

96
POOL EVAPORATION MODEL
Water diluted mixtures (1)

How to deal with water diluted mixtures?

1. Create diluted mixture
• Effects of solubility not considered
• Binary interaction (polar bindings) may lead to non-Ideal behaviour (≠ Raoult’s law)

2. Calculate evaporation rate of the pure substance

3. Use that evaporation rate for the dispersion modelling but simulate as a mixture (so toxic properties are
also adapted)

However, this results tend to be unreliable when there are strong polar chemicals present in the mixture!!

97
POOL EVAPORATION MODEL
Water diluted mixtures (2)
How to deal with water diluted strong polar chemicals (e.g. HCl-water, HNO3-water, etc.)?
If liquid vapour pressure deviates from ideal mixture….
1. Derive partial vapour pressure at Tstorage from experimental values or literature (≠ molar fraction * vapour pressure)
And either:
2. Adjust Tpool to corresponding temperature (read from graph “Liq. vapour pressure vs T” of the pure substance)
+ Adjust Tsubsoil and Tambient to obtain correct ∆T between pool and surroundings
or Adjust Antoine’s equation parameters A, B, C, D to describe correct vapour pressure (requires curve fitting app)
+ Adjust Tboiling
3. Simulate the pool evaporation with pure substance and adjusted temperatures

The dispersion calculation should be calculated with pure substance too, and the adjusted Tambient should be restored to its
original value. Also, toxic properties should be adapted to those of the mixture.

98
EXERCISE POOL EVAPORATION:
EXERCISE 6

99
LESSONS FROM POOL EVAPORATION EXERCISES

• Boiling liquid → heat transfer from soil to pool

→ higher evaporation rate for pool in water (water is constantly refreshed,  ΔT)

• Non-boiling liquid → solar heat radiation

• Presence of a bund → Pool size has big influence (kg/s∙m2 is determining)

100
 FIRE MODELS IN EFFECTS

02.12.2022
FIRE MODELS IN EFFECTS
Fire models

Combustion &
Gas fireball (direct
BLEVE fireball Jet fire Pool fire toxic combustion
ignition)
products

Also…
• Delayed ignition flash fire → “Dispersion: Flammable cloud model” (determines flammable contour - LFL)

• Direct flame contact → 100% lethal = > 35 kW/m2

• Damage translation → using Probit relations (1st degree, 2nd degree, 3rd degree lethal burns)

102
POINTS OF ATTENTION
BLEVE fireball:
• Dynamic (recommended) or static model
• Cold or hot BLEVE (fire underneath / external impact)
• Has overpressure effect: “BLEVE Blast model” included
Gas Fireball: NEW!

• Immediate ignition due to instantaneous release of compressed gas cloud (not drifting)
• Replaces former “Flash Fire” model (simple model that did not calculate radiation damage).
• Calculates both heat radiation and overpressure results
Jet fire:
• Connect to any gas / 2-phase outflow model to obtain flow rate
• Dominating influence of height and outflow angle
Pool fire:
• Free shape (circular, polygon, rectangle, rim fire)
Solid flame radiation approach
• “Yellow book” or “Two-zone pool fire” (HSE) model

103
BLEVE FIREBALL
The BLEVE phenomena is one of the most feared consequences for LPG transport

Korea BLEVE: short movie

104
BLEVE FIREBALL
User can choose between two models
• Dynamic (recommended)
• Based on the theory of Martinsen & Marx
• More realistic
• Time dependent
• Describes growing and rising of the fireball

• Static
• Based on the theory of the Yellow Book
• Simple modelling
• Not time dependent
• Fireball occurs at a static height

105
BLEVE FIREBALL
Model functionality
• Two types of BLEVE
• Cold (external impact) → Burst pressure = Pvapour
• Hot (fire underneath) → Burst pressure > Pvapour (or Tliquid ↑)

• Overpressure effect is included → “BLEVE Blast model”

Overpressure effects <<< heat radiation (fireball) effects

• Enter volume with storage conditions (instead of “mass in vessel”)

• It calculates 3 * adiabatic flash (BLEVE mass not always full inventory)

106
GAS FIREBALL NEW!

Direct ignition
• Based on the direct ignition of an instantaneous release of compressed gas.

• Replaces the former “Flash Fire (Instantaneous Gas Release) model”, which assumed
• Purple Book simple methodology
• Assumed no gas expansion due to combustion
• Did not include rising and growing behaviour of the fireball
• Assumed hemispheric expansion of the gas to UFL
• Only within this footprint, 100% lethality was assumed
• Not applicable for hydrogen (as it underestimated fire dimensions)
• Did not calculate heat radiation nor overpressure contours

107
GAS FIREBALL NEW!

Direct ignition

• The new Gas Fireball model calculates:

• footprint of the expansion phenomena.
• heat radiation and overpressure results
• rising and growing behaviour of the fireball

108
EXERCISE BLEVE FIREBALL:
EXERCISE 7

109
JET FIRE
EFFECTS chooses between two models
• Chamberlain
• Solid flame shape model
• For jet fires due to pure vapour phase releases
• For vertical/inclined releases
• Flame represented as a frustum

• Cook
• Adaptation of Chamberlain model to allow the calculation
of jet fires due to two-phase releases

110
JET FIRE
Input parameters
• Mass flow rate → Link to the representative release rate of a gas or two-phase outflow model

• Fraction of the flame covered by soot → the more soot the less heat radiated
• = 0 → No soot, clear flame
• = 1 → completely covered flame by soot
• Pool fires tend to generate more soot than jet fires due to the larger chain of hydrocarbons
• For jet fires we usually assume good mixing with air at the boundaries of the flame
• Therefore, usually there is no soot.
• Exception: hydrogen (as we expect a lower SEP than predicted)

• Maximum heat exposure duration → typically 20 seconds (time that takes the receiver to seek for shelter)

111
JET FIRE
Modelling challenge: fireball / vertical jet

The first seconds of an ignited jet fire might Propane Fireball

be modelled as a fireball! jet

a) Condensation, propane T = - 43 °C
b) Ignition, looks like a fireball (1 s)
c) Rising and growing of the fireball
d) Stable situation → Jet fire
Fireball vertical
rises
jet
+ jet flame
flame

112
JET FIRE
Latest improvement: hydrogen jet fires
• Characteristics H2 jet fire
• Wide flammable range (between 4% and 75% in air)
• Low ignition energy
• Burns quickly
• Low radiant heat
• Almost invisible

• More realistic approach

• SEP forced to 70 kW/m2 for pure H2 releases → corresponds to almost invisible flame radiating in the far infrared.

• Be aware that CH4/H2 mixtures have a much higher SEP (>>70 kW/m2).

113
POOL FIRE
The user can choose between two models
• Yellow Book model
• Simple model

• Rew & Hulbert (HSE) model

• Recommended model
• More realistic approach
• Distinguishes clear flame (bottom) from sooty flame (top), which have dedicated SEP values
• SEP, soot fraction and fraction of combustion heat radiated taken from substance specific table

114
POOL FIRE
Non-circular pools
POOL FIRE
Non-circular pools

• Pool fire model allows to “draw” shape of pool:,

“circular”, “rectangular”, “rim fire” or “polygon” → allows realistic
pool shape determined by bunds, tank pit or ditches.

• Flame shape determined by: Geometry + wind direction

• Footprints presented: Heat radiation & lethality

• DON’T try to draw holes for remaining tanks:

• Radiation is determined by “outside”
• Non burning area is relevant for flame height

116
POOL FIRE: REW & HULBERT Max. surface Unobscured ratio =
Substance specific values (1) Max. burning rate emissive power 1 - soot

• Calculation of the burning rate

If chemical listed in table…
mሶ b = mሶ "∞ ∙ 1 − e−D∙k∙β
If chemical not listed in table…
• Burgess relation
• Soot fraction = 0.8
• Fraction heat radiated = 0.35

• k∙β
To correct the reduction of burning rate with
decreasing diameter. The larger the diameter
the close the burning rate to the value in the
table.
117
POOL FIRE: REW & HULBERT Max. surface Unobscured ratio =
Substance specific values (2) Max. burning rate emissive power 1 - soot

• SEPmax
Heat emission through clear lower part of
flame.

• km
Extinction factor with which transparency of
the flame is corrected. For sooty flames and
large diameters this factor can be canceled.

• UR
Unobscured ratio corresponding to 1 – soot.
E.g. UR = 0.02 → 98% flame protected by soot.
→ 2% radiation remains in lower
part 118
EXERCISE POOL FIRE:
EXERCISE 8

119
FLASH FIRE
Delayed ignition

• Based on the delayed ignition of a dispersing cloud → Calculated with the “flammable cloud dispersion
model”.

• The model can’t calculate the heat radiation footprint because the duration of the flash fire is unknown.

• Calculates footprint based on LFL concentration.

120
COMBUSTION & TOXIC COMBUSTION PRODUCTS

• Model to calculate the formation rate of toxic combustion products due to warehouse fires.

• The warehouse material can be introduced as either a liquid or a solid.

• The model can be used as the previous calculation required to run the dispersion model “plume rise
from fires”

NOx
CO HCl HF
Soot2 HBr
SO2
121
FIRE MODELS: CONSEQUENCES DUE TO HEAT RADIATION
• Effects to humans due to heat radiation damage:

Expressed as “Dose”. Calculated with Probit function dependent on heat load (Q) and time (t):

Dose = Q4/3 ∙ t → translated into 1st, 2nd and 3rd degree burns

• Lethality

If Q < 35 kW/m2 → safe inside / If Q > 35 kW/m2 → 100% lethal / Direct flame contact → 100% lethal

• Clothing protection: Clothing (= protection) = 0.14* Lethality

Note that clothing burning may occur although lethality = 0

• Damage relations in Green book, PGS 1

122
LESSONS FROM FIRE EXERCISES

• BLEVE:
• Hot or Cold ? → determined by pressure at failure
• Static or Dynamic → Dynamic accounts for rising and growing of the fireball
• Heat radiation (and overpressure results)

• Jet fire
• Influence outflow angle

• Pool fire:
• Differences between with/without clothing protection
• Burning rate, soot formation (2-zone has more realistic values)

123
 DISPERSION MODELS IN EFFECTS

02.12.2022
DISPERSION MODELS IN EFFECTS
SMOKE PLUME GAS PLG LIQUID
Combustion model Gas release Two-phase release Liquid release Liquid cascade

Combustion
products & Rainout rate
heat generated
Spray release Pool evaporation
Airborne mass rate

Plume rise Dispersion

from fires

Concentration Concentration Flammable cloud Toxic Dose

Also…
• Damage translation for toxic models → Dosage and lethality outdoors (grid in map)
→ Dosage and lethality indoors (grid in map) 125
POINTS OF ATTENTION
Dispersion: dilution of airborne material
Gas turbulent free jet NEW!
• Under expanded jet creates huge turbulence Now incorporated within ALL
dispersion models
• Calculates rapid initial dilution due to turbulence
Dispersion (NEW)
• To simulate lighter than air, dense gas and neutral gas dispersion
• Concentration: at time t after the release, maximum concentration versus distance
• Toxic dose: cumulated (inhaled) toxic dose. Toxic dose = Cn ∙ t.
• Flammable cloud: explosive mass in cloud (above LEL), top view area and enclosed volume of the LEL cloud.
Neutral gas
• Only for “pure passive” situations! (no influence impulse, density differences, droplets)
Plume rise from fires (NEW)
• To simulate plume rise phenomenon due to warehouse fires
• Can be connected to the (toxic) combustion products fire model
126
GAS TURBULENT FREE JET
Why is needed?
Mechanisms of a gas turbulent free jet:
1. Upon the release of a ↑↑P gas, a jet is formed due its high velocity.
2. Velocity inside jet >> Velocity surrounding air
3. Friction is generated between high velocity gas and surrounding air
Calculates the jet’s predilution due
4. High velocity induces strong turbulence. to the turbulence created during a
5. Turbulence causes air entrainment and forces jet to spread sideways high-speed jet release

Mass outflow D (at limit of momentum)

Explosive mass (for dispersion)

6. This creates fast mixing / diluting with air at boundary jet

7. Some potential explosive mass already lost → vapour cloud diluted below LFL

127
GAS TURBULENT FREE JET
Integration within dispersion models
The gas turbulent free jet model has been integrated within all dispersion
models and it is activated for jet releases regardless of whether the
dispersion’s behaviour is dense, neutral or light.

Why?
• Eliminates the risk of neglecting the gas turbulent free jet phase
• Allows a better connection between the TFJ and the dispersion models
• Requires less inputs and additional steps from the user
• Some extensions have been implemented in the dispersion models to
better accommodate the linking to the TFJ model
All dispersion models as of v12 simulate the transition from turbulent
region to expanded jet without the user having to take any additional
steps.

128
GAS TURBULENT FREE JET NEW!

Improved high-speed releases

Why?
• Hydrogen is commonly stored and transported at very high pressures: 200-900 bar
• Gaseous highly-pressurized releases lead to under-expanded high-speed jet outflow
• Which leads to generation of greater turbulence upon release
• High velocity jet leads to the cloud’s rapid dilution before we having buoyancy driven or passive
dispersion

What has been improved?

• Turbulent free jet model more accurately accounts for under-expanded supersonic (high-speed)
releases
• The turbulent free jet model is integrated within the new Dispersion Model.
DISPERSION: DENSE, LIGHT, NEUTRAL GAS BEHAVIOUR
Which one to expect and why?
• Dense
→ Liquid droplets in cloud (e.g. spray release, instantaneous flashing liquid)
→ Extremely cold cloud (e.g. LNG)
→ Heavy molecules in pure vapour (e.g. chlorine)

• Neutral → Standard situations (e.g. non boiling pool evaporation of a gasoline leak,…)

• Light → Buoyant chemical substances (e.g. hydrogen, ammonia, methane gas,…)

130
ONE DISPERSION MODEL NEW!

To simulate all dispersion behaviours

Why?
• With the energy transition comes the use of new energy carriers (hydrogen & ammonia)
• Hydrogen is very buoyant
• Ammonia often released as refrigerated pressurized liquefied gas
• Dense behaviour upon release due to cloud containing cold liquid droplets
• Transition to lighter than air behaviour due to mixing with air and heat exchange

• The former “dense gas dispersion model” in v11 was originally developed for dense gas applications, and
was never validated against experimental data for lighter than air cases.

• That is why we needed a dispersion model that would simulate these buoyant plumes

131
DISPERSION MODEL
To simulate all dispersion behaviours
The Dispersion Model takes into account the entire physics of the dispersion process and accounts for density
differences, making it suitable for:

It also properly simulates possible transition behaviour such as: from dense to lighter than air, or from dense to
passive, …

132
DISPERSION MODEL
To simulate the entire physics of the dispersion

New Dispersion Model

originally based on theory of
SLAB but with improvements:
• For dense gas dispersion
modelling
• Validated for neutral gas
dispersion
• Extended to account for
the physics of buoyant jets
and plumes
DISPERSION MODEL
v11 vs v12

v11
• Dense gas dispersion model Different mathematical approaches depending on
• Neutral gas dispersion model dispersion behaviour

v12
• ONE Dispersion Model to simulate
Benefits:
• Dense gas
• Lifting plume behaviour included
• Lighter than air gas
• Automatic dispersion behaviour
• Neutral gas
• No discontinuities
• (Ad-hoc neutral gas dispersion model)
DISPERSION MODEL
What about combined models?

Note:
• Combined models will only use the new
Dispersion model

• Neutral gas dispersion model will be

available as ad-hoc
EXAMPLES OF DISPERSION
Instantaneous Dense Gas Dispersion (1)

136
EXAMPLES OF DISPERSION
Instantaneous Dense Gas Dispersion (2)

137
EXAMPLES OF DISPERSION
Continuous Dense Gas Dispersion

Effect of wind speed

138
INPUTS IN DISPERSION MODEL
Input parameters: Type of release

• Instantaneous → dispersion due to catastrophic rupture of a vessel

• We need to know the amount of mass that has been released
• For liquefied gas releases, link to Liquefied Gas Instantaneous Release model to calculate amount of
mass airborne

• Evaporating pool release → dispersion due the evaporation from a liquid spill
• Link to Pool Evaporation model to calculate the evaporation rate from a liquid spill

139
INPUTS IN DISPERSION MODEL
Input parameters: Type of release

• Horizontal and vertical jet release → dispersion due to jet release

• If gas jet release: link to Gas release model
• If liquefied gas jet release: link to Liquefied Gas Bottom Discharge + Spray release model
• The Gas Turbulent Free Jet model is incorporated within this type of dispersion calculation (for dense,
light or neutral)

• Low momentum release → for the rare occasions where you have a jet release released at very low speed
• to reproduce experimental conditions
• site with a lot of pipes and high-speed jet release bounces against surrounding geometries. Then
momentum is gone, and gas expands. Low momentum release can be used to simulate dispersion
after impingement occurs.

140
NEUTRAL GAS DISPERSION MODEL (AD-HOC MODEL)
When to use it?
• The new Dispersion Model substitutes the former neutral gas dispersion model.
• Neutral gas dispersion model remains available as an ad-hoc model
• When to use?
• For “pure passive” situations
• No impulse
• No density differences
• No droplets
• Simplified approach based on a Gaussian Plume Model
• Recommendation → use the new Dispersion Model for all dispersion behaviours
• Mathematical approach in the new dispersion model is based on the description of the entire physics
of the dispersion phenomenon

141
NEUTRAL GAS DISPERSION MODEL
Input parameters: Type of (semi-)continuous source
Type of (semi-)continuous source for Neutral Gas Dispersion (NGD)
• Evaporating pool release → to simulate the dispersion from an evaporating liquid spills
The model automatically adjusts the initial source dimensions of the dispersion based on provided pool surface
area

Lz = σz Lx = Radius pool Ly = Diameter pool

(accounts for initial
predilution)

• Jet release → NGD model will automatically calculate the Gas Turbulent Free Jet zone
• User defined window → To apply NGD model in pure passive initial conditions (e.g. Lee Wake situation)

142
NEUTRAL GAS DISPERSION MODEL
Input parameters: Source dimensions for user defined window
• Initial source dimensions for user defined window in Neutral Gas Dispersion
• Length source in wind (x) direction
• Length source in crosswind (y) direction wind
• Length source in vertical (z) direction

• Type of release
• Instantaneous → initial dimensions correspond to a CUBE of gas
• Semi-continuous → outflow through virtual outflow WINDOW of gas

• Source dimensions, imply a pre-dilution: Cinitial < 100 vol%

• By default: 0 x 0 x 0.
• Corrected to be consistent with density of 100 vol% chemical at ambient conditions.
143
ALL DISPERSION MODELS
Input parameters: Atmospheric stability (1)

Plume behaviour function of atmospheric stability conditions.

a) Unstable (“looping”)

b) Neutral (“coning”)

c) Stable above and below source: surface inversion (“fanning”)

d) Stable below source only: surface inversion (“lofting”)

e) Stable above source only: elevated inversion (“fumigation”)

144
ALL DISPERSION MODELS
Input parameters: Atmospheric stability (2)
• Pasquill/Gifford produced empirical results for variations of σy and σz for typical atmospheric turbulence
conditions
Class Definition Wind speed Day, Day, Day, Night, Night,
[m/s] Strong sun Moderate sun Slight sun overcast or > Clear or < 3/8
A Very unstable
4/8 low cloud cloud
B Unstable
2 A A-B B - -
C Lightly unstable 2-3 A-B B C E F
D Neutral 3-5 B B-C C D E
E Stable 5-6 C C-D D D D
F Very stable 6 C D D D D

• Turbulence defined by combination of = Pasquill class + wind speed + subsurface roughness length
• Monin-Obukov method: another way of defining atmospheric stability, using M-O length and mixing layer
height

145
ALL DISPERSION MODELS
Input parameters: Surface roughness length

• Surface roughness length → Describes additional turbulence due to friction at surface

• Defined classes for the roughness of the terrain, z0
• More obstacles = More friction = More turbulence = More mixing → Lower concentrations
Number Roughness length Description
1 0.0002 m Open water, at least 5 km
2 0.005 m Mud flats, snow, no vegetation
3 0.03 m Open flat terrain, grass, few isolated objects
4 0.1 m Low crops, occasional large obstacles, x/h > 20
5 0.25 m High crops, scattered large objects, 15 < x/h < 20
6 0.5 m Parkland, bushes, numerous obstacles, x/h < 15
7 1.0 m Regular large obstacles coverage (suburb, forest)
8 3.0 m City centre with high- and low rising buildings

See Table 44 from the Purple Book

146
ALL DISPERSION MODELS
Input parameters: Concentration averaging time

• Dispersion concentration averaging time → Rule of thumb: Toxic = 600 s / Flammable = 20 s

147
ALL DISPERSION MODELS
Input parameters: Others

• Weather class → atmospheric stability, wind speed, air temperature, etc.

• Initial liquid mass fraction (for dense gas) → calculated with:

Liquefied Gas Spray Release Model (semi-continuous)
Liquefied Gas Instantaneous Release Model (instantaneous)

→ dispersion model contains a subroutine to calculate evaporation from

liquid droplets which influences on T of cloud & heat exchange.

148
TOXIC DISPERSION
Input parameters: Toxic Exposure duration time
“Maximum toxic exposure duration” based on…
• Release duration → Non-limited exposure. Receiver is exposed for the whole duration of the release
• Cloud exposure → The exposure starts when cloud arrives to the receiver
• Time until sheltering → The exposure stops when people reacts & look for shelter
Concentration vs. Time for a 600 sec release at various distances

300 mtr 600 mtr 1200 mtr 1800 mtr 2400 mtr
Exposure duration = “max. duration release”

1,200,000
Concentration [mg/m3]

1,000,000 Exposure duration =

“time until sheltering”
800,000

600,000
Exposure duration = “max. exposure cloud”
400,000

200,000

0
0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 3,000 3,250 3,500 3,750
Time [s] 149
TOXIC DISPERSION
Input parameters: Indoor lethality

Lethality indoors = 10% of outdoor lethality → can be very underestimating:

Dose inside highly affected by

1. Ventilation: refreshment rate
2. Outside concentration profile: max level, duration cloud

150
FLAMMABLE DISPERSION
Input parameters: Reporting time flammable cloud

“Reporting time flammable cloud”. The explosive mass, length and width of the cloud is calculated based on…

• Time maximum area cloud → tMAC → Results at the time when cloud has reached its max. area
• Cloud is bigger in size but more diluted (less explosive mass)

• Time maximum explosive mass → tMEM → Results at the time when cloud has reached its max. explosive mass
• Cloud is smaller in size but more concentrated (more explosive mass)

• User defined → Results at a time defined by the user

151
FLAMMABLE DISPERSION
Input parameters: Reporting time flammable cloud

LFL at tmem LFL at tmac

LFL outer contour

152
 ALL DISPERSION MODELS
Output parameters: Dynamic cloud presentation

• Flammable model → Colours of dynamic contour: based on LFL concentration

• Toxic model → Colours of dynamic contour: based on 1% threshold lethal concentration
• Concentration model: Colours are based upon lowest requested threshold to present (AEGL/IDLH/VRW…)

153
LIMITATIONS INTEGRAL DISPERSION MODELS

• Influence of large obstacles on concentrations can be big (not included) → Dispersion concentrations are
averaged out

• Short scale modelling requires detailed 3D definition surroundings: important CFD input = geometry
→ Typical application field of Gexcon FLACS, FLACS Fire, FLACS Risk

• However, CFD…
• High complexity, long calculation duration
• Results only valid for one wind-direction, stability-class, …

154
LIMITATIONS INTEGRAL DISPERSION MODELS

• Atmosphere is stochastic in nature → meteorological conditions are dynamic in reality

• A dispersion experiment is always only ONE realisation of the stochastic nature of the atmosphere

• Results from large scale experiments are not reproducible

• Wind speed, direction and stability will vary with location, height and time:

Dispersion models applicable to maximum 10 km !!

155
LIMITATIONS INTEGRAL DISPERSION MODELS

Experiments vs model prediction

• Dispersion models are estimation of reality

• Stochastic meteorology effects are averaged out

Possible solution: for flammable clouds sometimes choose 50% LFL

as concentration contour

156
 EXERCISES DISPERSION:
REAL WORLD EXAMPLES.
CASE 1 & CASE 2.

157
 NEW!
PLUME RISE FROM FIRES
• Model to predict
behaviour of (toxic)
smoke plumes that
occur upon a
combustion process:
• Pool fire
• Warehouse fire

• Will there be
pollutants at ground
CO2 HCl level?
HF
Soot HBr
SO2 NOx

158
PLUME RISE FROM FIRES NEW!

Modelling approach

• Start with combustion model to find fumes production & composition

• Select which combustion product to trace (NOx , SO2 , HCl, …)

*Since 2021, any chemical/combustion product can be monitored
Oil warehouse fire near Haifa’s port (Sept ‘19)

• Define weather conditions and wind direction

159
PLUME RISE FROM FIRES
Inversion layer
PLUME RISE FROM FIRES NEW!

Modelling approach

Atmospheric stability is taken into account

• Tendency of atmosphere to resist or enhance vertical motion Night T
profile
• Presence of a temperature inversion layer is taken into account in

Increasing altitude
the dispersion behaviour

Increasing temperature
Top of inversion
75 m

Afternoon T
50 m profile

Increasing altitude
25 m

Stable atmosphere Unstable atmosphere

Increasing temperature
161
PLUME RISE FROM FIRES NEW!

Presence of a mixing layer

Penetration of and reflection against the mixing layer

Sideview shows “penetration” or potential “reflection” of the mixing layer height

162
PLUME RISE FROM FIRES NEW!

Limitations:
• Heat production dominates behaviour (uncertain, oxygen restricted fire?)
• No plume meandering (important at x >> 10 km)
• No deposition

Since 2021, any chemical/combustion product can be monitored

163
EXERCISE PLUME RISE FROM FIRES:
CASE 3.

164
 EXPLOSION MODELS IN EFFECTS

02.12.2022
EXPLOSION MODELS IN EFFECTS
Explosion
models

Solid explosion Vapour Cloud Explosion Rupture of

BLEVE Blast Gas Blast
(TNT) (Multi Energy) Vessels

Also…
• Damage translation → Fraction of mortality indoors vs distance
→ Fraction of mortality outdoors vs distance
→ Eardrum damage vs distance
→ Probability that glass is broken vs distance

166
POINTS OF ATTENTION
TNT model:
• For detonation of solids
• Also used for VCE but requires equivalency factor → Less suitable for VCE.
Multi Energy model:
• For (Confined) Vapour Cloud Explosion (VCE)
BLEVE Blast model:
• For BLEVE overpressure or vessel rupture (due to BLEVE)
Gas Blast model:
• For instantaneous release of gas (both direct & delayed ignition). Simulates depressurization phenomenon.
Rupture of Vessels model:
• For failure of HP vessels
• Includes throwing range of fragments
Dust explosions are not available yet! → determination depends on materials (grain, wood, flour,…), particle size, etc.

167
TNT MODEL
Methodology

• The TNT model is a blast curve method → TNT equivalency method

• Based on detonation of solids (short duration, high amplitude)

• Requires “equivalency factor”

“to relate the explosive power of an accidental explosion to
an equivalent TNT-charge”.
• Low reactive → 5%
• Medium reactive → 10%
• High reactive → 15%

168
MULTI ENERGY MODEL
Methodology

• The Multi Energy model is a blast curve method developed by TNO

• Developed for simulation of VCEs

• Requires estimation “Blast curve” + “% confined explosive mass”

• Results determined by
• Chemical properties
• Curve number
• % confined

169
MULTI ENERGY MODEL
Selection of blast curve (1)
• Set of 10 graphs (curve 1-10), overpressure versus distance P(x)
• Curve 1-2 → Very weak deflagration

• Curve 3-4 → Weak deflagration

• Curve 5 → Medium deflagration

• Curve 6-7 → Strong deflagration

• Curve 8-9 → Very strong deflagration

• Curve 10 → Detonation
170
 MULTI ENERGY MODEL
Selection of blast curve (2)
Curve number determines the (initial) overpressure, accounting for: Ignition strength, Obstruction & Parallel plane confinement
Blast strength Ignition strength Obstruction Parallel plane Curve
category (High / Low) (High / Low / None) (Confined / Unconfined) number

1 H H C 7 – 10
2 H H U 7 – 10 Conclusion:
3 L H C 5–7
Application of Multi Energy
4 H L C 5–7
5 H L U 4–6 model requires expert
6 H N C 4–6 judgement!
7 L H U 4–5
(Blast curve, % Confinement)
8 H N U 4–5
9 L L C 3–5
10 L L U 2–3
11 L N C 1–2
12 L N U 1

171
MULTI ENERGY MODEL
Extension with GAME relations (1)
To estimate the fraction of the cloud captured in congested area…

LFL footprint
Only small part of cloud
covers installations: e.g.
10%

The GAME model calculates % confinement and curve number based on overlap of “drifting cloud” with congestion area:
1. Define congestion area (receiver)
2. Link to flammable cloud dispersion model

172
MULTI ENERGY MODEL
Extension with GAME relations (2)
• Parameters of congestion block:
• 2D or 3D confinement → 2D (confined between parallel planes) vs 3D (confined between planes + ceiling)
• Volume Blockage Ratio (VBR) → amount of free space (not occupied by obstacles) available for expansion of a
burning cloud
• Typical obstacle diameter (Davg) → representative diameter for obstacles in congested area

• GAME describes max. overpressure as function of quantifiable parameters:

Pmax = 0.84 (VBR * Lp / Davg)2.75 . SL2.7 . Davg0.7 (open 3D)
Pmax = 3.38 (VBR * Lp / Davg)2.25 . SL2.7 . Davg0.7 (2D confinement)

173
BLEVE BLAST

• To calculate BLEVE overpressure or vessel rupture (due to BLEVE)

• Calculates 1% lethality contour and hazard distance to a threshold overpressure level

• Does not include heat radiation results. For that → BLEVE Fireball model

• The BLEVE Fireball model also contains the overpressure results as calculated in the BLEVE Blast model.

174
GAS BLAST

• To simulate the depressurization that occurs upon an instantaneous gas release

• Reports overpressure damage, calculates 1% lethality contour and hazard distance to a threshold
overpressure level

• The model is based on the “Rupture of Vessels” model which is used where the cause of vessel rupture is
“pressure vessel burst ideal gas”.

• Does not include heat radiation results. For that → Gas Fireball model

• The Gas Fireball model also contains the overpressure results as calculated in the Gas Blast model.

175
RUPTURE OF VESSELS

• To calculate failure of highly pressurized vessels

• Includes throwing range of fragments (calculates maximum range of fragments)

• Different causes for vessel rupture depending on type of burst are included:
• Pressure vessel burst (with ideal or non-ideal gas)
• BLEVE phenomenon due to rupture of vessel containing pressurized liquefied gas
• Decomposition of materials → requires mass of decomposable chemical & heat of reaction
• Runaway reaction (with ideal or non-ideal gas)
• Internal explosion

176
OTHER VCE MODELS

Numerical simulations → CFD based: FLACS

• Requires detailed 3D geometry
• Total site: multiple sources/receivers/ignition points
• Multiple wind angle’s: laborious & expensive

177
 RECEIVERS IN EFFECTS

02.12.2022
RECEIVERS
• Receiver = area that interferes with a consequence model. Can be added to end-models (i.e. fire, explosion & dispersion)

• Two types:
• Vulnerable Areas → Areas with specific vulnerability thresholds
• Congestion Areas → Congestion region for VCE evaluation (to be linked to Multi Energy model)

• Definition → Node = Receivers: <Right mouse> <Add>

• Definition of wind direction → “Worst case” / “Defined by model”

• IMPORTANT: Dynamic concentration presentation = YES (we need a time depending cloud location)

179
EXERCISE: DRAWING RECEIVERS (AREAS)
Drawing shapes on top of a background map

1. Press <Edit> button: mouse cursor

2. Start pinpointing coordinates of shape
3. Red dots = active dots → drag to move, click to delete, click outside to add
4. Press <Edit> to finish drawing

180
VULNERABLE AREAS

• Specific vulnerability thresholds → heat radiation, overpressure or toxic

• Specific definition of population indoors → number of people, fraction indoors
• Use in combination with end-models → i.e. fire models, toxic dispersion, overpressure
• Results specific for area → nr fatalities, max heat radiation/overpressure

181
CONGESTION AREAS
• A flammable cloud will only give overpressure IF a degree of congestion/confinement occurs (Flash fire vs VCE)
• Better estimation of cloud confinement and explosive strength with definition of
• Area and height → cloud overlap and maximum enclosed volume
• Level of congestion → fixed blast strength or GAME parameters
• Used only in combination with flammable cloud dispersion models
• Results specific for area: explosive mass, blast curve strength

182
CONGESTION AREAS
Congestion area parameters & confined mass → determine Blast strength

183
 EXERCISE EXPLOSION MODELS & RECEIVERS:
CASE 4: METHANE VCE

EXERCISE TANK OVERFILLING MODEL

CASE 5: PENTANE VCE DUE TO LIQUID CASCADE

184
 DAMAGE MODELS IN EFFECTS

02.12.2022
TRANSLATION FROM PHYSICAL EFFECT TO DAMAGE
The Probit function (Pr) is a measure of probability (fraction of people) to assess damage due to a physical effect:
𝐏𝐫 = 𝐀 + 𝐁 ∙ 𝐥𝐧 𝐃𝐨𝐬𝐞
Where dose can be
• Toxic dose, D = Cn ∙ t, with C in [mg/m3 or ppm]
• Heat load, D = Qn ∙ t, with Q in [W/m2]
• Pressure impulse, D = Pn ∙ t, with P in [bar or Pa] (Bevi uses 100 & 300 mbar as 2.5% & 100% lethality)

Points of attention:
• A, B, n Probit values are substance/effect specific & unit dependent !
• Accepted values can be country specific!
• Changing or adding Probit values per Chemical: stored in “user defined database”

NOTE: See “Mortality/Probit calculator” to show relation between C, t, % lethality

186
TRANSLATION FROM PHYSICAL EFFECT TO DAMAGE
% 0 1 2 3 4 5 6 7 8 9
Probit = 2,67 → 1% damage 0 - 2,67 2,95 3,12 3,25 3,36 3,45 3,52 3,59 3,66
10 3,72 3,77 3,82 3,897 3,92 3,96 4,01 4,05 4,08 4,12
Probit = 5,00 → 50% damage
20 4,16 4,19 4,23 4,26 4,29 4,33 4,36 4,39 4,42 4,45
Probit = 8,09 → 99% damage
30 4,48 4,50 4,53 4,56 4,59 4,61 4,64 4,67 4,69 4,72
40 4,75 4,77 4,80 4,82 4,85 4,87 4,90 4,92 4,95 4,97
Damage is calculated with…
50 5,00 5,03 5,05 5,08 5,10 5,13 5,15 5,18 5,20 5,23
1 Pr − 5
p = ∙ 1 + erf 60 5,25 5,28 5,31 5,33 5,36 5,39 5,41 5,44 5,47 5,50
2 2
70 5,52 5,55 5,58 5,61 5,64 5,67 5,71 5,74 5,77 5,81
80 5,84 5,88 5,92 5,95 5,99 6,04 6,08 6,13 6,18 6,23
90 6,28 6,34 6,41 6,48 6,55 6,64 6,75 6,88 7,08 7,33

-- 0,0 0,1 0,3 0,3 0,4 0,5 0,6 0,7 0,8 0,9
99 7,33 7,37 7,41 7,46 7,51 7,58 7,65 7,75 7,88 8,09
187
DAMAGE RELATIONS IN EFFECTS MODELS
Heat radiation & Toxic effects

• Toxic dose, D = Cn ∙ t, with C in [mg/m3] and t in [min]

• Lethality of Acrylonitrile → Pr = -8,6 + 1 ∙ ln(C1.3 ∙ t)
• Lethality of Chlorine → Pr = -6,35 + 0,5 ∙ ln(C2,75 ∙ t)
• Chemical database contains Probit values (A, B, n) for some toxic chemicals

• Heat load, D = Qn ∙ t, with Q in [W/m2] and t in [s]:

• Lethality Green Book → Pr = -36,38 + 2,56 ∙ ln(Q1.33 ∙ t)
• 1st degree burns → Pr = -39,83 + 3,0186 ∙ ln(Q1.33 ∙ t)

188
DAMAGE RELATIONS IN EFFECTS MODELS
Overpressure

• Pressure impulse, D = Pn ∙ t, with D in [Pa] and t in [s]. No consensus for Pressure-Damage relation.
Methods:
• Pressure Impulse Probit (Flemish region )
• Pr = -8,23 + 1,31 ∙ ln(P1 ∙ t)
• Pressure vs Lethality relation (Singapore)
• 5 psi → 3% lethality
• 7 psi → 10% lethality
• 10 psi → 50% lethality
• Threshold Pressure Level (The Netherlands)
• 300 mbar → 100% lethality (total destruction level)
• 100 mbar → 2,5% lethality

189
DAMAGE CALCULATIONS IN EFFECTS MODELS
The EFFECTS models include multiple approaches to translate from consequence effects to damage:
• Toxic exposure models
• Fraction of mortality vs distance
• Toxic dose vs distance
• Dosage indoors/outdoors for every grid location

• Heat radiation models

• 1st, 2nd, 3rd (lethal) degree burns vs distance

• Overpressure models
• Fraction of mortality indoors/outdoors vs distance
• Eardrum damage vs distance
• Breaking of windows vs distance

190
DEDICATED DAMAGE MODELS
Damage models

Explosion damage Toxic damage

Consequences Consequences
Conversion from Protection by Consequences
to structures to individuals
animal experiments remaining indoors to individuals
Direct effects Indirect effects
During a During a semi-
Construction Glass continuous release continuous release
fragments fragments

Heat radiation damage

Consequences Consequences to Heat load to reach Time to reach a

to individuals population groups critical T (for structures) critical T

For steel For glass 191

DEDICATED DAMAGE MODELS
These damage models are hardly used because damage relations are already implemented in the EFFECTS models.
• Explosion damage
• Consequences to structures → Calculates damage to brick houses, damage to American-style houses, etc.
• Consequences to individuals → Calculates lung damage, body displacement, eardrum damage, etc.

• Heat radiation damage

• Consequences to individuals → Calculates 1st,2nd, 3rd degree burns taking into account clothing effects .
• Consequences to population → It also takes into account the total amount of people exposed.

• Toxic damage
• Conversion animal experiments → To derive Probits (A, B, n) from LC50 (Animal type) experimental values
• Protection by remaining indoors→ Calculates lethality indoors/outdoors given an outdoor concentration
• Consequences to individuals → To calculate lethality due to toxic inhalation

192
 COMBINED MODELS IN EFFECTS

02.12.2022
COMBINED MODELS
The easy way
Combined models contain the full chain of models required to calculate all possible phenomena for a specific
LoC scenario.

Combined models

GAS LoC Liquid LoC

G1: Instantaneous G2: Release 10 min G3: Leak G1: Instantaneous G2: Release 10 min G3: Leak

Liquefied Gas LoC

Unified LoC

G1: Instantaneous G2: Release 10 min G3: Leak

194
POINTS OF ATTENTION
Unified LOC:
• Universal model
• Detects chemical phase depending on storage conditions
• Contains all the “phase specific LOCs”

Phase specific LOC:

• The user shall choose the chemical phase: gas, liquefied gas or liquid.

Type of combined LoC scenario:

• G1 → Instantaneous release due to the vessel’s catastrophic rupture.
• G2 → Release of all the inventory in 10 minutes. Calculates the corresponding hole size.
• G3 → Leak though a hole which size is specified by the user.

195
POINTS OF ATTENTION

Phenomena to calculate:
• Auto detect mode → EFFECTS assesses relevant phenomena based on substance properties and conditions
• User defined → Only BLEVE, only Toxic etc.

Expert vs Simple mode

• The user should know the independent models well → “Expert mode” contains approx. 95 input parameters
• Change to “Simple mode” to reduce to 20 input parameters

Choice of dispersion model

• Always new dispersion model (as it simulates dense, light and neutral behaviour)
• Neutral gas dispersion model is no longer used in combined models

196
POINTS OF ATTENTION

2 types of dispersion models in the combined models:

• Flammable Cloud: to assess the amount of flammable mass that can potentially be involved in an explosion

• Toxic Cloud: to assess the toxic concentration of the cloud and dose received depending on the exposure duration

197
ADVANTAGES OF USING THE COMBINED MODELS
Source rate for the dispersion calculation of a two-phase release:
• Air borne spray → leads to dense gas dispersion
+
• Evaporation rate (from the liquid pool due to rain-out) → leads to neutral gas dispersion

The two dispersion calculations are shown: Pool (plume mode) vs Instantaneous (puff mode)

Calculation of the direct ignition of pool for continuous releases (G2 & G3):
• Max. size of the pool is determined by equilibrium between release rate & burn rate

For substances that are toxic & flammable…

• Only one representative release rate is considered

198
EXERCISE COMBINED MODELS:
• Start the Unified LOC (Loss Of Containment),
• Use simple mode, click defaults (Propane, Instantaneous)
• Evaluate results
• Change Chemical to “Ammonia”: Recalculate and evaluate

199
MODELS THAT REQUIRE CUMULATION OF SOURCE RATES
Mode: Instantaneous
Gas LoC G1 2 dispersion models (Toxic cloud, Flammable cloud)

Mode: semi-continuous
Gas LoC Gas LoC G2 2 dispersion models (Toxic cloud, Flammable cloud)

Mode: semi-continuous
Gas LoC G3 2 dispersion models (Toxic cloud, Flammable cloud)

Mode: Combination
Liq gas LoC instantaneous & semi-continuous
4 dispersion models
G1

LoC liq gas

Unified LoC Liq gas LoC
G2
Mode: Combination semi-continuous
Liq gas LoC 2 dispersion models

G3

Liquid LoC
Mode: Instantaneous
G1 2 dispersion models (Toxic cloud, Flammable cloud)

Liquid LoC
Liquid LoC Mode: semi-continuous
G2 2 dispersion models (Toxic cloud, Flammable cloud)

Liquid LoC 200

Mode: semi-continuous
G3 2 dispersion models (Toxic cloud, Flammable cloud)
CONTINUOUS TWO-PHASE RELEASES (G2/G3)
Model tree
• Source rate for the dispersion calculation = Airborne mass from the spray + Pool evaporation rate
• If pure “dry” gas: pressure and hole dimensions required (turbulent mixing jet)

Flammable
Airborne mass cloud
Combine
Outflow Spray

Rainout
Temperature sources
Toxic cloud
Pool size
Pool-
Wind speed Evaporation rate
evap
Chemical

201
CONTINUOUS TWO-PHASE RELEASES (G2/G3)
Cumulation of source terms

• Mass flowrate → mcombined = mspray + mevaporation

mspray
• Liquid fraction → Φcombined = Φspray ∙
mcombined
mspray ∙ Tspray + mevaporation ∙ Tevaporation
• Representative Release duration → Tcombined =
mcombined

Spray mspray Combine mcombined

TPDIS Φcombined
Release mass rates

mevaporation
Tcombined

Dispersion model:
Toxic Cloud or Flammable Cloud
Pool
evaporation

202
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Model tree
• Dispersion simulation is based on a PUFF (instantaneous) or a PLUME (semi-continuous)

Flammable
Cloud
Instantaneous mode: PUFF

Toxic Dose
Airborne mass Cumulate
Flash
sources
Rainout

Flammable
Cloud
Pool size Pool Semi-continuous mode: PLUME
evaporation
Wind speed Toxic Dose

203
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Cumulation of source terms

Instantaneous mode: PUFF

Liquefied Gas Airborne mass (Mflash) Instantaneous

Instantaneous Release dispersion
Rainout Cumulate
mevaporation
Pool Continuous
Evaporation dispersion

Semi-continuous mode: PLUME

204
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Cumulation of flammable clouds

Cumulate flammable (explosives):

1. Determine dominating cloud size (between instantaneous & continuous dispersion)

2. Cumulate explosive mass at tmac

3. Correct LEL cloud size dominating source with ratio Mcumulated / Moriginal

205
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Cumulation of flammable clouds

What is the dominating cloud (at time tmac)? → In this case, instantaneous flash presents a dominating
flammable cloud.
Session 37 - Explosive mass (HGDE Instantaneous 27)
Session 37 - Explosive mass (HGDE Poolevap 31) Session 37 - Area (HGDE Instantaneous 27) Session 37 - Area (HGDE Poolevap 31)

16,000 140,000
15,000
130,000
14,000
120,000
13,000
110,000
12,000

Area (HGDE Instantaneous 27) [m2]

100,000
11,000
90,000
10,000

9,000 80,000

8,000 70,000

7,000 60,000
6,000
50,000
5,000
40,000
4,000
30,000
3,000
20,000
2,000

1,000 10,000

0 0
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Time [s] Time [s]
206
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Cumulation of toxic clouds

Cumulate toxics:

1. Calculate toxic dose grid for both models (instantaneous & continuous dispersion)

2. Cumulate the toxic dose grids (cumulated dose as function of x and y)

3. Calculate lethality from cumulated dose grid: (lethality as function of x and y)

207
INSTANTANEOUS TWO-PHASE RELEASES (G1)
Cumulation of toxic clouds
Combining → cumulation of the 2 calculated doses (instantaneous – flash vs semi-continuous – pool)

Pressurized ammonia Refrigerated ammonia

208
OBJECTIVES TRAINING EFFECTS

❑ Understand the approach of performing consequence calculations

❑ Understand and distinguish several modes of releases, and potential phenomena occurring (which model to use for
which situation)

❑ Understand the approach of performing a consequence calculation

❑ Be able to interpret and analyse the results of consequence calculations

❑ Be able to use and apply the software tool EFFECTS 11

209
UPCOMING UPDATES (2023-2024)
Software improvements
• Import/Export tool

Model extensions
• 3D results for fire, dispersion and explosion
• Directional shielding
• Liquid jet fires
• Improvements on mixtures: water dilution, binary interaction, liquid compressibility, etc.
• Dispersion model that handles 3D terrain and mitigating measures (e.g., walls)

210
FINALIZING REMARKS

• Question during usage → e-mail helpdesk [email protected]

• Feedback is essential to us (either bug reports, comments on usage, or request for functionalities):
• We learn from your reactions!
• The programme improves

• You’ll be informed about updates and new releases (also see: download.gexcon.nl)

• Any remaining questions?

211
EFFECTS & RISKCURVES QUICKSTART TRAINING

Any remaining questions?

https://quickstart.gexcon.nl/ (Password: SafetySoftware)

212