RiskInformed Separation Distances for Hydrogen Fueling Stations Jeffrey

Risk-Informed Separation Distances for Hydrogen Fueling Stations Jeffrey La. Chance Sandia National Laboratories 10/21/2021 Preliminary Data

Outline • • 10/21/2021 Project Background Separation Distances Risk-Informed Approach Description Application to an Example Facility Input Data and Assumptions for QRA Models Preliminary Results Conclusions Future Work Preliminary Data

Project Background • Work performed under U. S. DOE Hydrogen, Fuel Cells & Infrastructure Technologies Program, Multi-Year Research, Development and Demonstration Plan – Hydrogen Safety, Codes & Standard R&D • Sandia National Laboratories is developing the scientific basis for assessing credible safety scenarios and providing the technical data for use in the development of codes and standards – Includes experimentation and modeling to understand behavior of hydrogen for different release scenarios – Use of Quantitative Risk Assessment (QRA) methods to help establish separation (setback, safety) distances at hydrogen facilities and to identify accident prevention and mitigation strategies for key risk drivers 10/21/2021 Preliminary Data

Separation Distances • Specified distances in codes separating H 2 components from the public, structures, flammable material, and ignition sources • Distance vary with possible consequences from hydrogen releases (e. g. , radiation heat fluxes or overpressures) • Distances influenced by facility design parameters (e. g. , hydrogen pressure and volume), available safety features (e. g. , isolation valves), and release parameters (e. g. , leak size and location) Separation distances based solely on the consequences of hydrogen leaks can be large! 10/21/2021 Preliminary Data

Example of Consequence-Based Separation Distances for a Jet Fire Leak Diameter (mm) 10/21/2021 Preliminary Data

Separation Distances for Different Consequence Measures Consequence Measure 10/21/2021 Preliminary Data

Risk-Informed Approach • Uses risk insights plus other considerations to help define code requirements Risk = Frequency X Consequence from all accidents • Requires hydrogen component leak frequencies • Requires definition of important consequences • Uses QRA and consequence models to evaluate risk • Requires definition of acceptable risk levels • Accounts for parameter and modeling uncertainty present in analysis 10/21/2021 Preliminary Data

Risk Approach for Establishing Separation Distances Cumulative frequency Cumulative of accidentsfrequency requiring of accidents requiring this separation distance 10/21/2021 Preliminary Data

Use of Risk Eliminates Large Leaks from Consideration Pipe Leak Frequency Increasing Leak Diameter 10/21/2021 Preliminary Data

Application to Example Facility • Currently evaluating risk-informed separation distances for a representative fueling facility – To demonstrate risk methodology – To evaluate important facility features (e. g. , gas volume and leak isolation features) – To determine importance of modeling parameters (e. g. , data, geometry, temporal effects) – To identify key risk scenarios and identify possible ways to reduce the risk to acceptable levels • Work presented is focused on hydrogen jet releases from gas pipes and gas storage cylinders 10/21/2021 Preliminary Data

Example Facility Description • Facility can refuel 100 cars/day • All components located outside • Liquid hydrogen facility with no-onsite production – Liquid storage volume of 3000 kg @ 0. 55 Mpa – Cryo pump (11 L/min) used to provide pressurized liquid to vaporizer when pressure in gas storage decreases • Natural draft vaporizer • Gas storage was sized for 500 kg of hydrogen(12. 63 m 3 for 10000 psig facility) – Three separate cascades • Two refueling dispensers • Facility meets codes and standards 10/21/2021 Preliminary Data

Leakage Frequency Distributions 10/21/2021 Preliminary Data

Consequence Parameters and Risk Criteria Used in Current Analysis Consequence Parameters – Radiant Heat Flux from Jet Fires: • 1. 6 k. W/m 2 – no harm to individuals for long exposures • 4. 7 k. W/m 2 – injury (second degree burns) within 20 seconds • 25 k. W/m 2 – 100% lethality within 1 minute; equipment and structural damage – Hydrogen Concentration from Un-ignited Releases: • 4%, 6%, and 8% concentrations – lower flammability limit • Risk Criteria – Frequency of Fatality to Individual at Separation Distance Used as Upper Bound Accident Frequency Criteria • <2 E-4/yr – fatality risk from all other high-risk hazards in society 10/21/2021 Preliminary Data

Gas Pipe Analysis Assumptions • Leakage can occur in 50 m of 13. 5 mm diameter pipe, 4 valves, 2 instrument lines (4. 23 mm D), and 3 flanges • Hydrogen leak or flame detector sends signal to isolation valve resulting in closure within 10 s • Immediate ignition results in fatality at separation distance if not automatically detected and isolated (no credit for manual detection and isolation) • Delayed ignition of un-isolated gas jet results in flash fire – Fatality assumed out to distance corresponding to LFL (used to determine separation distance) • Pipe leak orientation was assumed directed at lot line resulting in maximum separation distances 10/21/2021 Preliminary Data

Gas Pipe Leak Event Tree Pipe Leak or Rupture Downstream of PIPE_LEAK 10/21/2021 pipe leak - (New Event Tree) Immediate Ignition Detection of Automatic Isolation Delayed Ignition of of Hydrogen Jet Hydrogen or Flame of Pipe within 10 s Hydrogen I-IGNITION DETECTION ISOLATION Preliminary Data D-IGNITION # END-STATE-NAMES 1 JET-FIRE-(10 -S) 2 JET-FIRE 3 JET-FIRE 4 GAS-RELEASE-(10 -S) 5 FLASH-FIRE 6 GAS-RELEASE 7 FLASH-FIRE 8 GAS-RELEASE 2007/01/27 Page 0

Gas Pipe Results: Un-isolated Jet Fires Mean frequency of any size un-isolated pipe leak < 1 E-6/yr 10/21/2021 Preliminary Data

Gas Pipe Results: Isolated Jet fire Exposure time to isolated jet fires is short which reduces potential for structural or equipment damage and personnel injury. 10/21/2021 Preliminary Data

Frequency of Fatality from Isolated Jet Fire (10 s exposure) 10 s exposure to 25 kw/m 2 heat flux results in 20% probability of a fatality (reducing the exposure time to 5 s would further reduce the probability to 5%). Operating Pressure 10/21/2021 Preliminary Data

Gas Pipe Results: Flash Fires Delayed ignition assumed to result in flash fire and fatality out to LFL. Hydrogen Concentration 10/21/2021 Preliminary Data

Gas Storage Analysis Assumptions • For each storage cylinder, leakage can occur in 1 m of 4. 23 mm diameter pipe and 1 valve attached to the cylinder, and from the cylinder itself. Leakage can also occur in manifold and 4 attached valves. • No protective barriers are installed around the gas storage area • No isolation of a gas storage leak from these components is possible • Gas storage is outdoors and thus there is little potential for a hydrogen detonation • Immediate ignition results in fatality at separation distance if not automatically detected and isolated (no credit for manual detection and isolation) • Delayed ignition of un-isolated gas jet results in flash fire – Fatality assumed out to distance corresponding to LFL (used to determine separation distance) • Leak orientation was assumed to result in worst consequences (sensitivity study performed) 10/21/2021 Preliminary Data

Gas Storage Leak Event Tree 10/21/2021 Preliminary Data

Gas Storage Results: Un-isolated Jet Fires Risk-informed separation distances are affected by leakage contribution from different components. Leakage from all components Manifold and cylinder leaks Heat Flux Cylinder leaks 10/21/2021 Preliminary Data

Gas Storage Results: Flash Fires Flash fires, not jet fires, require the longest separation distances (36 m for an LFL of 4% vs. 18 m for 25 kw/m 2 heat flux - 2 E-4/yr fatality criteria). Hydrogen Concentration 10/21/2021 Preliminary Data

Gas Storage Sensitivity Study – Volume of Stored Gas Limiting gas storage volume can lead to reduced risk-informed separation distances. Mass of Gaseous Hydrogen 10/21/2021 Preliminary Data

Gas Storage Sensitivity Study – Size of Gas Cylinders Increasing the gas cylinder size reduces the leakage frequency. Cylinder Size 10/21/2021 Preliminary Data

Example Separation Distances: Lot Lines Maximum Risk-Informed Separation Distances (m) Pipe Leaks 1 Gas Storage Leaks 2 Criteria 5000 psig 2 E-4/yr 0 0 0 5 E-5/yr 0 0 1 E-5/yr 9 5 E-6/yr 12 1. 2. 10/21/2021 10000 psig 15000 psig 13 -26 16 -32 19 -36 0 17 -30 22 -44 24 -49 13 15 29 -59 38 -76 44 -87 17 20 40 -72 40 -82 46 -92 Pipe separation distances are limited by isolated jet fires (25 Kw/m 2). Gas storage separation distances are limited by flash fires (8% - 4% H 2 concentration). Preliminary Data

Conclusions from Example Pipe Analysis • Properly designed and installed hydrogen and flame detection and automatic isolation capability can reduce the risk from piping leaks – Frequency of un-isolated jet fire and flash fire scenarios can be small enough to justify short risk-informed separation distances. – Separation distances are limited by isolated jet fire scenarios. – Isolation capability at dispensers, vaporizers, and a facility with a compressor will likely result in similar results as above. 10/21/2021 Preliminary Data

Conclusions from Example Gas Storage Analysis • Gas storage leakage events are likely the dominant risk contributors for the example facility (accidents involving liquid hydrogen have not yet been evaluated) – Relatively short risk-informed separation distances (<20 m) could be justified for jet fires for a fatality risk criteria >1 E-5/yr. – However, risk-informed separation distances are limited by flash fire scenarios (separation distances for high pressure facilities >40 m for a fatality risk criteria <1 E-5/yr). – Reducing the gas storage volume or using larger volume cylinders can reduce the frequency of gas leaks and resulting risk-informed separation distances. – Risk-informed separation distances may still be too long for gas storage releases. Barriers surrounding the storage area may be required to reduce the risk to acceptable levels. 10/21/2021 Preliminary Data

Future Efforts • Continue evaluating safety distances for example facility – Liquid hydrogen storage leaks/ruptures – Evaluate alternate methods for pressurizing gas (cryo pump vs. compressor) – Dispensers • Evaluate facilities using different methods for onsite hydrogen production (gas reforming and electrolysis) • Improve methodology including consideration of time-dependent impacts, geometry factors, and incorporation of uncertainty • Get consensus on failure data for use in QRA (e. g. , leak frequencies and component failure rates) • Identify range of models or data for accident phenomenology probabilities (e. g. , ignition and detection probabilities) • Identify key risk drivers for hydrogen facilities and identify what can be done to reduce the risk and separation distances to acceptable levels • Extend evaluation to other types of hydrogen facilities 10/21/2021 Preliminary Data