quantitative risk analysisto guide stationdesign ... · first-of-its-kind integration platform...
TRANSCRIPT
Quantitative Risk Analysis to Guide Station Design
Gabriela Bran-Anleu Team Chris LaFleur Alice Muna Brian Ehrhart Myra Blaylock Ethan Hecht Sandia National Laboratories
SAND2018-10661 PE September 11 2018 This presentation does not contain any proprietary confidential or otherwise restricted information
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia LLC a wholly owned subsidiary of Honeywell International Inc for the US Department of Energyrsquos National Nuclear Security Administration under contract DE-NA-0003525
il i $
457$m$
$
Quantitative Risk Assessment is enabling infrastructure deployment
Lot$Line$(73$m)$
Serv ce$Sta3on$ A5endant$Bui d ng
(91$x$152$m)$
366$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$207$m$$
Gasoline$Dispensers$
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
2005 2007 2009 2011 2013 2015 2017
Risk assessment proposed for
hydrogen systems at ICHS
QRA applied to indoor refueling to inform code revision
Established risk-informed processes for separation distances
Public release of HyRAM RampD tool
QRA-informed separation distances
in NFPA 2 20 station penetration potential due to QRA
ISO TC197 WG24 incorporating QRA and behavior modeling
Performance-based system layout demonstrated
PLL 5084e-04 FAR 01161 AIR 2322e-06
$149$m$$
2
RW Schefer et al International Journal of Hydrogen Energy 31 (2006) 1332 ndash 1340 1337
Fig 4 Visible IR and UV images of turbulent hydrogen-jet flamedjet = 794 mm
in the visible and IR images is about 55 m in the ver-tical direction and about 35 m in the UV image Theirregularities in the outer edges of the flame reflect theunsteady turbulent mixing of the fuel with ambient airThe UV camera exposure was gated for 160 s using theintensifier This exposure time is sufficiently short tonearly freeze the flow motion and reveal many featuresof the instantaneous flame structure The chemilumi-nescence intensity recorded within each flame image isspatially irregular and also varies from image to imagewhich reflects the temporal and spatial variations foundin the instantaneous structure of turbulent flames
Flame lengths based on all three images were used todetermine the time-average flame length (Fig 5) Theaverage flame length was then taken as the flame lengthaveraged over five successive frames around the indi-cated time for each point The flame length decreaseswith time due to the decrease in mass flow rate astank pressure is reduced It can be seen that the short-est flame lengths are based on the UV flame emissionwhile the longest flame lengths are based on IR emis-sion The average values for L visL IR and L uvL IR areabout 088 and 078 respectively As discussed previ-ously it is expected that L UV should indicate the loca-tion of the primary reaction zone where the fuel is be-ing oxidized while L IR should be more indicative of thehigh temperature combustion products The measuredflame length ratios are consistent with this proposedflame behavior
Based on an analysis of the transition frommomentum-controlled to buoyancy-controlled turbu-
00
10
20
30
40
50
60
0 20 40 60 80 100
Lir (m)
Lvis (m)
Luv (m)
Flam
e Le
ngth
(m)
Time (sec)
Fig 5 Flame length history using visible infrared and ultravioletflame emission
lent jet flame dynamics Delichatsios [6] developeda useful correlation for turbulent flame lengths Thecorrelation is based on a non-dimensional Froude num-ber that measures the ratio of buoyancy to momentumforces in jet flames Using the nomenclature of Turns[5] the Froude number is defined as
Fr f = uef32s
( e infin)14[( TfTinfin)gdj ]12 (4)
where ue is the jet exit velocity fs is the mass fractionof fuel at stoichiometric conditions ( e infin) is the ra-tio of jet gas density to ambient gas density dj is thejet exit diameter and Tf is the peak flame temper-ature rise due to combustion heat release Small val-ues of Fr f correspond to buoyancy-dominated flameswhile large values of Fr f correspond to momentum-dominated flames Note that the parameters known tocontrol turbulent flame length such as jet diameter andflow rate stoichiometry and ( e infin) are included inFr f Further a non-dimensional flame length L lowast canbe defined as
L lowast = L visfs
dj ( e infin)12 = L visfs
dlowast (5)
where L vis is the visible flame length and dlowast is the jetmomentum diameter (=dj ( e infin)12) Fig 6 showsthe resulting correlation of flame length data fromRef [3] for a range of fuels (H2 C3H8 and CH4) and
Hydrogen behavior studies are at the foundation of consequence modeling capabilities
2005 2007 2009 2011 2013 2015 2017
Advanced laser diagnostics applied to turbulent H2 combustion
Ignition of under-expanded H2 jets
Buoyant jet flame model with multi-source radiation
Laboratory-scale characterization of LH2 plumes and jets
Barrier walls for risk reduction
Radiative properties of H2 flames quantified
Experiment and simulation of indoor H2 releases
Ignition limits of turbulent H2 flows
3
Enab methods data tools for hydrog fety
Service$S on$en an uil ing$
li $ is s rs$
$
Coordinated activities to enable consistent rigorous and accepted safety analysis
ling en sa
Develop and validatescientific models to accurately predicthazards and harm from liquid releases
flames etc
Behavior RampD (SCS 010)
Develop integrated methods and algorithms for enabling
consistent traceable and rigorous QRA
Risk RampD (SCS 011)
Apply QRA amp behavior models to real problems in hydrogen
infrastructure and emerging technology
Application in SCS (SCS025)
Lot$Line$(73$m)$
ta3 A5 d t$B d
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$ $207$m$$
Gaso ne D pen e
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
Developing methods data tools for H2 safety amp SCS 4
Building a Scientific Platform for Hydrogen QRA
2 System amp hazarddescription
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
5
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
il i $
457$m$
$
Quantitative Risk Assessment is enabling infrastructure deployment
Lot$Line$(73$m)$
Serv ce$Sta3on$ A5endant$Bui d ng
(91$x$152$m)$
366$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$207$m$$
Gasoline$Dispensers$
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
2005 2007 2009 2011 2013 2015 2017
Risk assessment proposed for
hydrogen systems at ICHS
QRA applied to indoor refueling to inform code revision
Established risk-informed processes for separation distances
Public release of HyRAM RampD tool
QRA-informed separation distances
in NFPA 2 20 station penetration potential due to QRA
ISO TC197 WG24 incorporating QRA and behavior modeling
Performance-based system layout demonstrated
PLL 5084e-04 FAR 01161 AIR 2322e-06
$149$m$$
2
RW Schefer et al International Journal of Hydrogen Energy 31 (2006) 1332 ndash 1340 1337
Fig 4 Visible IR and UV images of turbulent hydrogen-jet flamedjet = 794 mm
in the visible and IR images is about 55 m in the ver-tical direction and about 35 m in the UV image Theirregularities in the outer edges of the flame reflect theunsteady turbulent mixing of the fuel with ambient airThe UV camera exposure was gated for 160 s using theintensifier This exposure time is sufficiently short tonearly freeze the flow motion and reveal many featuresof the instantaneous flame structure The chemilumi-nescence intensity recorded within each flame image isspatially irregular and also varies from image to imagewhich reflects the temporal and spatial variations foundin the instantaneous structure of turbulent flames
Flame lengths based on all three images were used todetermine the time-average flame length (Fig 5) Theaverage flame length was then taken as the flame lengthaveraged over five successive frames around the indi-cated time for each point The flame length decreaseswith time due to the decrease in mass flow rate astank pressure is reduced It can be seen that the short-est flame lengths are based on the UV flame emissionwhile the longest flame lengths are based on IR emis-sion The average values for L visL IR and L uvL IR areabout 088 and 078 respectively As discussed previ-ously it is expected that L UV should indicate the loca-tion of the primary reaction zone where the fuel is be-ing oxidized while L IR should be more indicative of thehigh temperature combustion products The measuredflame length ratios are consistent with this proposedflame behavior
Based on an analysis of the transition frommomentum-controlled to buoyancy-controlled turbu-
00
10
20
30
40
50
60
0 20 40 60 80 100
Lir (m)
Lvis (m)
Luv (m)
Flam
e Le
ngth
(m)
Time (sec)
Fig 5 Flame length history using visible infrared and ultravioletflame emission
lent jet flame dynamics Delichatsios [6] developeda useful correlation for turbulent flame lengths Thecorrelation is based on a non-dimensional Froude num-ber that measures the ratio of buoyancy to momentumforces in jet flames Using the nomenclature of Turns[5] the Froude number is defined as
Fr f = uef32s
( e infin)14[( TfTinfin)gdj ]12 (4)
where ue is the jet exit velocity fs is the mass fractionof fuel at stoichiometric conditions ( e infin) is the ra-tio of jet gas density to ambient gas density dj is thejet exit diameter and Tf is the peak flame temper-ature rise due to combustion heat release Small val-ues of Fr f correspond to buoyancy-dominated flameswhile large values of Fr f correspond to momentum-dominated flames Note that the parameters known tocontrol turbulent flame length such as jet diameter andflow rate stoichiometry and ( e infin) are included inFr f Further a non-dimensional flame length L lowast canbe defined as
L lowast = L visfs
dj ( e infin)12 = L visfs
dlowast (5)
where L vis is the visible flame length and dlowast is the jetmomentum diameter (=dj ( e infin)12) Fig 6 showsthe resulting correlation of flame length data fromRef [3] for a range of fuels (H2 C3H8 and CH4) and
Hydrogen behavior studies are at the foundation of consequence modeling capabilities
2005 2007 2009 2011 2013 2015 2017
Advanced laser diagnostics applied to turbulent H2 combustion
Ignition of under-expanded H2 jets
Buoyant jet flame model with multi-source radiation
Laboratory-scale characterization of LH2 plumes and jets
Barrier walls for risk reduction
Radiative properties of H2 flames quantified
Experiment and simulation of indoor H2 releases
Ignition limits of turbulent H2 flows
3
Enab methods data tools for hydrog fety
Service$S on$en an uil ing$
li $ is s rs$
$
Coordinated activities to enable consistent rigorous and accepted safety analysis
ling en sa
Develop and validatescientific models to accurately predicthazards and harm from liquid releases
flames etc
Behavior RampD (SCS 010)
Develop integrated methods and algorithms for enabling
consistent traceable and rigorous QRA
Risk RampD (SCS 011)
Apply QRA amp behavior models to real problems in hydrogen
infrastructure and emerging technology
Application in SCS (SCS025)
Lot$Line$(73$m)$
ta3 A5 d t$B d
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$ $207$m$$
Gaso ne D pen e
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
Developing methods data tools for H2 safety amp SCS 4
Building a Scientific Platform for Hydrogen QRA
2 System amp hazarddescription
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
5
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
RW Schefer et al International Journal of Hydrogen Energy 31 (2006) 1332 ndash 1340 1337
Fig 4 Visible IR and UV images of turbulent hydrogen-jet flamedjet = 794 mm
in the visible and IR images is about 55 m in the ver-tical direction and about 35 m in the UV image Theirregularities in the outer edges of the flame reflect theunsteady turbulent mixing of the fuel with ambient airThe UV camera exposure was gated for 160 s using theintensifier This exposure time is sufficiently short tonearly freeze the flow motion and reveal many featuresof the instantaneous flame structure The chemilumi-nescence intensity recorded within each flame image isspatially irregular and also varies from image to imagewhich reflects the temporal and spatial variations foundin the instantaneous structure of turbulent flames
Flame lengths based on all three images were used todetermine the time-average flame length (Fig 5) Theaverage flame length was then taken as the flame lengthaveraged over five successive frames around the indi-cated time for each point The flame length decreaseswith time due to the decrease in mass flow rate astank pressure is reduced It can be seen that the short-est flame lengths are based on the UV flame emissionwhile the longest flame lengths are based on IR emis-sion The average values for L visL IR and L uvL IR areabout 088 and 078 respectively As discussed previ-ously it is expected that L UV should indicate the loca-tion of the primary reaction zone where the fuel is be-ing oxidized while L IR should be more indicative of thehigh temperature combustion products The measuredflame length ratios are consistent with this proposedflame behavior
Based on an analysis of the transition frommomentum-controlled to buoyancy-controlled turbu-
00
10
20
30
40
50
60
0 20 40 60 80 100
Lir (m)
Lvis (m)
Luv (m)
Flam
e Le
ngth
(m)
Time (sec)
Fig 5 Flame length history using visible infrared and ultravioletflame emission
lent jet flame dynamics Delichatsios [6] developeda useful correlation for turbulent flame lengths Thecorrelation is based on a non-dimensional Froude num-ber that measures the ratio of buoyancy to momentumforces in jet flames Using the nomenclature of Turns[5] the Froude number is defined as
Fr f = uef32s
( e infin)14[( TfTinfin)gdj ]12 (4)
where ue is the jet exit velocity fs is the mass fractionof fuel at stoichiometric conditions ( e infin) is the ra-tio of jet gas density to ambient gas density dj is thejet exit diameter and Tf is the peak flame temper-ature rise due to combustion heat release Small val-ues of Fr f correspond to buoyancy-dominated flameswhile large values of Fr f correspond to momentum-dominated flames Note that the parameters known tocontrol turbulent flame length such as jet diameter andflow rate stoichiometry and ( e infin) are included inFr f Further a non-dimensional flame length L lowast canbe defined as
L lowast = L visfs
dj ( e infin)12 = L visfs
dlowast (5)
where L vis is the visible flame length and dlowast is the jetmomentum diameter (=dj ( e infin)12) Fig 6 showsthe resulting correlation of flame length data fromRef [3] for a range of fuels (H2 C3H8 and CH4) and
Hydrogen behavior studies are at the foundation of consequence modeling capabilities
2005 2007 2009 2011 2013 2015 2017
Advanced laser diagnostics applied to turbulent H2 combustion
Ignition of under-expanded H2 jets
Buoyant jet flame model with multi-source radiation
Laboratory-scale characterization of LH2 plumes and jets
Barrier walls for risk reduction
Radiative properties of H2 flames quantified
Experiment and simulation of indoor H2 releases
Ignition limits of turbulent H2 flows
3
Enab methods data tools for hydrog fety
Service$S on$en an uil ing$
li $ is s rs$
$
Coordinated activities to enable consistent rigorous and accepted safety analysis
ling en sa
Develop and validatescientific models to accurately predicthazards and harm from liquid releases
flames etc
Behavior RampD (SCS 010)
Develop integrated methods and algorithms for enabling
consistent traceable and rigorous QRA
Risk RampD (SCS 011)
Apply QRA amp behavior models to real problems in hydrogen
infrastructure and emerging technology
Application in SCS (SCS025)
Lot$Line$(73$m)$
ta3 A5 d t$B d
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$ $207$m$$
Gaso ne D pen e
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
Developing methods data tools for H2 safety amp SCS 4
Building a Scientific Platform for Hydrogen QRA
2 System amp hazarddescription
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
5
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Enab methods data tools for hydrog fety
Service$S on$en an uil ing$
li $ is s rs$
$
Coordinated activities to enable consistent rigorous and accepted safety analysis
ling en sa
Develop and validatescientific models to accurately predicthazards and harm from liquid releases
flames etc
Behavior RampD (SCS 010)
Develop integrated methods and algorithms for enabling
consistent traceable and rigorous QRA
Risk RampD (SCS 011)
Apply QRA amp behavior models to real problems in hydrogen
infrastructure and emerging technology
Application in SCS (SCS025)
Lot$Line$(73$m)$
ta3 A5 d t$B d
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classified$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$ $207$m$$
Gaso ne D pen e
Gasoline$Tanks$$ (Fill$and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
GH2
Developing methods data tools for H2 safety amp SCS 4
Building a Scientific Platform for Hydrogen QRA
2 System amp hazarddescription
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
5
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Building a Scientific Platform for Hydrogen QRA
2 System amp hazarddescription
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
5
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Challenge A quality QRA incorporates a large body of information from different areas
QRA data
Relevant hazards
Exposurescenarios Gas data System data Facility data Frequency
data Consequence
data Loss harm
data
Chemical properties
Component counts
Componentconfiguration
Warehouse configuration
Populationdata (human)
Populationdata (fuelcells)
Release occurrence
Componentfailures
Human errors
H2 release behavior
H2 dispersionbehavior
H2 accumulation behavior
Thermal effects
Overpressureeffects
Toxicityeffects
Accidents
Ignition occurrence
Mitigating event
occurrence
Fire [load]models
Jet flames
Explosiondeflagration[load] models
It is necessary tohellip bull Find best-available models amp data for all of these pieces
bull Validate those models bull And combine all into a single framework
6
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
HyRAM Making hydrogen safety science accessiblethrough integrated toolsFirst-of-its-kind integration platform for state-of-the-art hydrogensafety models amp data - built to put the RampD into the hands of industrysafety experts
Current release is version 1111341
Core functionalitybull Quantitative risk assessment (QRA)
methodologybull Frequency amp probability data for hydrogen
component failuresbull Fast-running models of hydrogen gas and
flame behaviorsKey featuresbull GUI amp Mathematics Middlewarebull Documented approach models algorithmsbull Flexible and expandable framework
supported by active RampD
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Mole Fraction andHeight of Flame
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
7
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Start
ChooseHyRAMMode Physics ModePhysics
Cut Set ProbabilitiesScenario Rankings (PLLContribution)
Importance MeasuresCalculations
Major Elements of HyRAM Software QRA Mode QRA
QRA Methodology bull Risk metrics calculations FAR PLL AIR bull Scenario models amp frequency bull Release frequency bull Harm models Generic Freq amp Prob data bull Ignition probabilities bull Component leak frequencies (9 types) Software Language bull C for GUI and QRA (planned
System Description User Inputs Number of components
QRAMode
Scenario User Inputs GasDetection Credit
Event Tree Embedded in HyRAM
DataProbability User Inputs componentleak size probabilities failure mode
probabilities
Fault Tree Embedded in HyRAM
DataProbability User Inputs Ignition Probabilities
ScenarioFrequency
All Five Leak Sizes Analyzed
Explosion Jet Fire
H2 Release
Leak Detected
Leak Isolated
Immediate Ignition
Delayed Ignition
Shutdown
No ignition
Jet Fire
Explosion
System Description Inputs Pipeouterdiameter H2 temperature and pressure
ambient temperatureand pressure
Run Jet Fire Model to calculate heat flux at occupant positions
Consequence Model- Harm InputsThermal probit model exposure time
Consequence Model Inputs Notional nozzle model radiative sourcemodel
Plot on Occupant Positions and Heat Flux
Consequence Model User Inputs Peak overpressureand impulse for each release
size
Calculate probability and consequence (fatalities) given exposure to overpressure
Consequence Model- Harm InputsOverpressure probit model
conversion of QRA to Python) bull Python for Physics Modules Documentation bull Algorithm report (SAND2017-2998) bull User guide (SAND2018-0749)
Calculate probability and consequence (fatality) given exposure to jet fireheat flux
System Description Inputs Number of vehicles number of fuelings per day
number of vehicle operating days facility dimensions number and placement of
occupants
System Description Inputs Random number generator inputs (random seed
exclusion radius)
Assign locations for each exposed person
Calculate annual frequency of each scenario (jet fire explosion) for each release size
Calculate probability and consequence (fatality) from all scenarios
Calculate risk metrics (PLL AIR FAR)
8
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angleUser Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode FlameTemperatureTrajectory
Flame TemperatureTrajectory Results
Jet FlameOverpressure
Overpressure
Physics
Choose JetFlame Model Flame Temperature
TrajectoryRadiative Heat Flux
RadiativeHeat Flux
User Input Ambient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat fluxpoints nozzle and radiative source models
Start
QRAQRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
OverpressureResults
Overpressure [kPa] Plot
FlammableMass[kg] Plot
Height ofFlammable Layerand Mole Fraction
Plot
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperatureand Distance
Temperature and Distance
3-D Heat Flux Map
Pressure LayerDepth and
ConcentrationHeat Flux atUser-Provided
Points
Mole Fraction and Height of Flame
Major Elements of HyRAM Software Physics Mode Physics models bull Properties of Hydrogen bull Unignited releases Orifice
flow Notional nozzles Gas jetplume Accumulation in enclosures
bull Ignited releases Jet flames
overpressures in enclosures
Software Language bull Python for Modules
bull C for GUI
Documentation bull Algorithm report
(SAND2017-2998) bull User guide (SAND2018-
0749)
ChooseHyRAMMode
User input Nozzle modelambient temperatureand
pressure H2 temperature andpressure leak diameter leak height fromfloor and release
angle User Input Ambient temperatureand pressure H2temperature and pressure tank volume leak
diameter discharge coefficients (orifice releaseblocking area) release distances dimensions andrelative location wind angle and speed ceiling and
vent dimensions and distances
Choose Model
Physics Mode Flame Temperature Trajectory
Flame Temperature Trajectory Results
Jet Flame Overpressure
Overpressure
Physics
Choose Jet Flame Model Flame Temperature
Trajectory Radiative Heat Flux
Radiative Heat Flux
User InputAmbient temperatureandpressure H2 temperature and pressureleak diameter and height from floor
relative humidity XYZ radiative heat flux points nozzle and radiative source models
Start
QRA QRAMode
User Output Options Output pressures at given timesmaximum time and pressure lines on plots
Overpressure Results
Gas Plume
Gas Plume
User Input Ambient temperatureand pressure H2temperature and pressure orifice diameter orifice
discharge coefficients and angle of jet
User Output Options Plot X ampY limits and contour line
Gas PlumeResults
Heat Flux Results
Temperature and Distance
Mole Fraction and Height of Flame
9
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Summary
bull HyRAM is an integration platform built to enable hydrogen safety for state-of-the-art H2 safety models ndash enables consistent industry-led QRA and consequence analysis with documented referenceable validated models
bull Demonstrated Impact Enabling the deployment of refueling stations by developing science-based risk-informed codes amp standards ndash Analyses for NFPA 2 and ISO TR-19880-1 ndash Benchmarked results (SAND2014-3416) Survey of proposed H2 stations
show that changes to NFPA 2 gaseous separation distance requirements increased station siting options by 20
10
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
$207$m$$
Service$S on$A5endant$Building$
asoline$ is ensers$
GH2$
Behavior RampD Risk RampD Application in SCS
Lot$Line$(73$m)$
ta3
(91$x$152$m)$
366$m$
457$m$
Building$ Openings$amp$
HVAC$ Intakes$ (73$m)$
Classifi ed$Electrical$(46$m)$
$165$m$$
$238$m$$
$149$m$$
G D p
Gasoline$Tanks$$ (Fill $and$Vent)$
Equipment$Dimensions$ 91$m$long$HP$storage$ 30$m$wide$TT$ 30$m$wide$HP$storage$
Thank you Gabriela Bran Anleu
Sandia National Laboratories gabranasandiagov
httpshyramsandiagov Research supported by DOE Fuel Cell Technologies Office
(DOE EEREFCTO)
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Technical Back-Up Slides
12
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Benefits of Reduced-Order Models bull Short run-time bull Modeling expert not required bull Useful for quantification
ndash If a hydrogen leak occurs how far away does the hazard get bull Useful for comparisons
ndash What is the effect on safety is a system size is reduced Reduced- Real Order Model System
13
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Greater Fidelity and Flexibility of QRA Models
Expand HyRAM QRA beyond Event Sequence Diagram fueling stations bull Customization of event and
fault trees bull Perform risk assessment and
calculate risk results in an efficient manner
Fault Trees bull Applicable for new hydrogen industries beyond fueling stations
bull Underlying physics-based analysis would remain the same
Customization of scenario will lead to broader application of HyRAM and hydrogen QRA
14
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
bull Validation of near-field model complete including mole fraction temperature and velocity
bull Development of diagnostic to measure full-scale cold vapor releases underway
bull Development of full-scale release experiments underway
15
Laboratory-scale characterization of LH2 plumes and jets
Validated LH2 release model will be used to risk-inform the revised LH2 bulk
separation
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
RampD provides science-based tools Examples of Scenario amp Probability modelsAccident sequencesbull Hazards considered Thermal effects (jet
fire) overpressure (explosiondeflagration)
Ignition probabilitybull Extrapolated from
methane ignition probabilities
bull Flow rate calculated using Release Characteristics module
Release frequency- Expected annual leak freq for each
component type -- Data developed from limited H2 data combined w data from other industries
Harm modelsbull Probability of fatality from exposure to heat
flux and overpressures ndash multiple options
2$amp(
= +- 012
3+ lowast 5( 78 +)
+ 5 $ ltlt=gt3(
lowast 3A0BC2(JetFire) = f(H2release) lowast (1minusPr(Detect)) lowast Pr(IgnImmed)
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000 8000
Thermal Dose ((kWm2)43 s)
Fata
lity
Eisenberg Tsao and Perry Lees TNO HSE Criteria
00
100
200
300
400
500
600
700
800
900
1000
10 100 1000
Peak Overpressure (kPA)Fa
talit
y TNO-Lung
TNO-HeadTNO-BodyTNO-Collapse
Risk~
C
N
(CN O ltCN)
16
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
0 50 100 150 2000
500
1000
1500
2000
2500
3000
3500
Axia
l dis
tanc
e (m
m)
Radial distance (mm)
1 of mean1 FFFlame Light up
RampD provides science-based tools Examples of Behavior amp Consequence modelsRelease Characteristicsbull Prediction of hydrogen jet
plumes (concentration boundaries)
bull Prediction of hydrogen jet flames
bull Simplified models of hydrogen sources (choked flow notional nozzles etc)
IgnitionFlame Light-upbull Prediction of ignition
(flammability factor concept)bull Identification of light-up
boundariesbull Prediction of sustained flame
Deflagration within Enclosuresbull Overpressure associated
with deflagration bull Quantitative role of
ventilation
Flame Radiationbull Flame integral model effects of buoyancybull Multi-source models significantly improve heat
flux prediction bull Surface reflection can be a major potential heat
flux contributor
17
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Overpressure amp layer modulesInput Release conditions and enclosure configuration
Output Overpressure (ignited) amp Height of accumulated layer (unignited)
bull Enables calculation of consequences inside of enclosures
bull Insight into enclosure design effectiveness of mitigations
18
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19
Building a scientific platform for hydrogen QRA
2 System amp hazard description
1 Set analysis goals
3 Cause analysis
4 Consequence analysis
5 Communicate Results
User-specific ndash Elicit from range of stakeholders
User-neutral ndash Establish science amp engineering basis (with user input)
Adding more flexibility for users
19