DOI: 10.3303/CET2290059 Paper Received: 28 January 2022; Revised: 22 March 2022; Accepted: 6 May 2022 Please cite this article as: Brzezińska D., Januszewski J., 2022, Enclosure safety in a case of hydrogen, CNG, or LNG accidental release from a bus tank, Chemical Engineering Transactions, 90, 349-354 DOI:10.3303/CET2290059 CHEMICAL ENGINEERING TRANSACTIONS VOL. 90, 2022 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Aleš Bernatík, Bruno Fabiano Copyright © 2022, AIDIC Servizi S.r.l. ISBN 978-88-95608-88-4; ISSN 2283-9216 Enclosure Safety in a Case of Hydrogen, CNG, or LNG Accidental Release from a Bus Tank Dorota Brzezińska, Janusz Januszewski Lodz University of Technology, Faculty of Process and Environmental Engineering dorota.brzezinska@p.lodz.pl, janusz.januszewski@dokt.p.lodz.pl Hydrogen, CNG and LNG are highly explosive gases, which can create extremely hazardous conditions if released into an enclosure. The gases’ behaviour was checked under accidental release conditions by CFD simulations. It has been proved that too large a gas release diameter orifice could create a dangerous situation, and a gas concentration level above the lower explosive limit, even before emergency ventilation activation. The suggested practical solution for the hazard mitigation could be a limitation of the gas installation diameters fitted on the buses. 1. Introduction New fuels industry development makes hydrogen and natural gas ideal for large-sized vehicles, like buses, which are more and more popular with the general public, due to their impressive potential as a source of energy with minimal environmental impact. The gases are seen as clean and as ecological fuels, which are mainly produced from renewable sources (Kaplan and Kopacz, 2020). They create new opportunities in the fuel market, together with difficult new challenges in the field of safety, mostly associated with enclosed garages or tunnels. It is reported that the increasing usage of alternative fuels in vehicles (electricity, hydrogen, LNG, LPG etc.) has significantly altered vehicle fire characteristics (Klassen and Ph, 2020). For example, hydrogen ignition energy is ten times lower than methane, and its flame velocity is eight times higher than that of methane (Kaplan and Kopacz, 2020). Also, lower explosive limits (LEL) of H2 and CH4, volumetric fractions equal 4% and 4.4%, respectively, are much lower than other flammable gases. Consequently, ensuring a safe infrastructure for such vehicles becomes an urgent goal. This goal could be reached by doing research on the gases’ dispersive, burning, and explosive phenomenon and protective measures mitigating hazards relative to gas-fuelled buses (Brzezińska, 2019). It was found that the release of gas fuel from the onboard storage tank could be a major risk, especially when initiated by an external fire. Such accidents could create different hazardous outcomes and are dependent on varying factors. The most dangerous of which, are situations where the release happens in an enclosed space, such as a garage. In such an instance, the most influenced elements would be, the enclosure volume, its ventilation, and hydrogen release rate (Molkov, 2012a), (Brennan, Makarov and Molkov, 2009), (Molkov, Shentsov and Quintiere, 2014), (Molkov et al., 2014). In the case of accidental gas release from the bus’s fuel tank installation, several scenarios are possible (Dorofeev, 2007), (Hajji, Bouteraa, et al., 2015), (Prasad, Pitts and Yang, 2011), (Brzezińska, Dziubiński and Markowski, 2017), and these can be captured in an event tree, as shown in Figure 1. As an initiating event, the release of the gas fuel from the bus’s system was taken into account, with ten different outcomes considered (Brennan and Molkov, 2013). The first possible scenario is its permeation through the tank walls; however, this phenomenon is prolonged and happens mainly in vehicles fuelled with liquid hydrogen, which is not popular (Molkov, 2012a). Another scenario could be the fuel release from the bus’s pipe installation, when the tank is closed by a safety valve. In this case, the gas spread could be much more rapid than previously thought possible, but its volume would be limited to the capacity of the pipe installation. The third scenario, which is taken into account most often in accidental gas fuel release analyses, is the gas spread directly from a tank through a thermally activated pressure relief device (TPRD), an obligatory element of the bus’s installation (Makarov, 349 mailto:janusz.januszewski@dokt.p.lodz.pl Shentsov, et al., 2018). This is also the most dangerous scenario and could happen in a case of the TPRD failure or a matter of fire, when the tank pressure increases due to surrounding temperature increase (Molkov, V.; Shentsov, 2016), (Rodionov, Wilkening and Moretto, 2011). The consequences of the scenarios described above could have different levels of severity, that mostly depend on, ignition sources existing in the enclosure, detection, and ventilation systems. The most dangerous outcome event possible could be a gas explosion, which could comprise of, a deflagration, detonation, or detonation-to- deflagration transition (DDT) (Molkov, 2012b), (Makarov, Hooker, et al., 2018), (Molkov and Bragin, 2015). This scenario could happen, when the gas cloud created in an enclosure, is in a concentration of at least above the lower explosive limit (LEL), at the same time a viable ignition source appears (Saffers and Molkov, 2014), (Najjar, 2013). This can occur when a ventilation system does not work (is not efficient enough/doesn’t exist/was not activated by detection system or its activation is too late) (Molkov, 2012a). This phenomenon is investigated in this paper. Another outcome event presented in the event tree (Figure 1) is the gas cumulation in the enclosure, which could also be the cause of another severe event. However, the most probable final event would be a slow gas spread outside of the enclosure. The next event, which is gas dilution and/or evacuation, doesn’t create dangerous situations in the enclosure and would happen when the protective measures, such as detection and ventilation system would be sufficiently effective, and/or the hydrogen release would not be sufficiently rapid enough (Weiner, 2014), (Hajji, Jouini, et al., 2015), (Tolias et al., 2019), (Brzezińska, 2021). Finally, one more severe outcome event to consider, would be a pressure peaking phenomenon (PPP), and this happens in enclosers under intensive lighter gas emission into a heavier gas, under minor ventilation conditions (Makarov, Shentsov, et al., 2018), (Molkov and Bragin, 2015), (Lach, Gaathaug and Vaagsaether, 2020). This scenario could be considered, especially in the case of hydrogen-fuelled vehicles, but because it happens very rarely, it is not investigated here. Figure 1. Event tree of gas fuel release accidental scenarios. The event tree in Figure 1 demonstrates how important and effective detection and ventilation systems would be in a case of a flammable gas release. But looking at this protective system in depth, it appears that its effectiveness could be inadequate if additional conditions are not realised, demonstrated by CFD simulation results of hydrogen and natural gas dispersion, within a confined space, under typical bus tanks release variations. The analyses confirm differences in explosive hazards in the case of hydrogen, CNG, or LNG release, in dependence on the tank orifice diameter, and proves that keeping the gas concentration in an enclosure below the lower explosive limit (LEL) by the time the ventilation system would be activated, could technically be possible if the release quantity is limited. This release limitation can be realized by using pipes in the gas installations with reduced diameters. The article also demonstrates how the gas dispersion phenomenon in an enclosure depends on, the gas type, tank pressure, and the orifice diameter, and in consequence – there are changes to the buoyancy-induced momentum and the diffusive motions ratio. The expected hydrogen and natural gas behaviour was projected by the Mach (M), Reynolds (Re), and Froud (Fr) numbers, which values predict its flow regimes (Brzezińska, 2021). The flow chart in Figure 2 presents subsequent steps of the analyses. Initiating event Release mode TPRD closing Immediate ignition Detection & Ventilation Delay ignition Outcome event Y 1. Gas dilution/evacuation Continued permeation from a tank Y 2. Deflagration (Detonation) N N 3. Gas cumulation Y 4. Jet fire (short life) Gas release Continuous release Y 5. Gas dilution/evacuation from a pipe instalation N N 6. Gas dilution/evacuation Y Failure release from Y 7. Jet fire car's instalation N Continuous release Y 8. Gas dilution/evacuation from a tank N Y 9. Deflagration (Detonation) N N 10. Preasure peaking phenom. 350 Event tree for derivation of possible incident scenarios The release regimes analysis CFD simulations Figure 2. The analysis steps. 2. The release regimes analysis Therefore, theoretical verification of the gas dispersion phenomenon could be realized on the basis of three familiar dimensionless numbers analysis: Mach (M), Reynolds (Re), and Froud (Fr), which in many publications are used for gas plume characterization (Krishnapisharody and Irons, 2013), (Fischer et al., 2013). The first step was a definition of the release regimes by the Mach and Reynolds numbers calculations. Where it was assumed that transition from laminar (buoyancy-controlled) diffusion jet to turbulent (momentum-controlled) starts to occur at Reynolds number of around Re = 2000, and choked sonic release are when M≥1 (Molkov and Saffers, 2013), (Prasad, 2014). The Mach and Reynolds numbers were calculated on the basis of equations (1) and (2), respectively (Andrzejewski et al., 2019). The Froud number indicated the influence of gravity on fluid motion, and this was used to evaluate how much buoyancy force affects hydrogen dispersion (Molkov, 2012a). The Fr~1 could mean an absence of the buoyancy effect and a lack of the diffusive distribution of the gas. The Froud number was calculated on the base of equation (3) (Molkov, 2012a). The basic gases’ and bus tanks’ parameters are presented in Table 1. The main differences between the tanks are their internal pressure and the difference in the specific gravity of the gases causing significant differences in their capacity. The significant differences in velocity at the orifices appear due to the different pressures. The calculated Reynolds, Mach, and Froud number values are presented in Table 2. Table 1. The gases’ and tanks’ parameters. Gas P Vol Mass LEL H2 350 0.312 7.5 4% vol (3,41x10-3 kg/m3) CNG 200 0.375 42.5 4.4% vol (2,88x10-2 kg/m3) LNG 15 0.750 345.0 4.4% vol (2,88x10-2 kg/m3) Table 2. Reynolds, Mach, and Froud number values for the accidental releases. No Gas V U D ρ0 μ M Re Fr S1 H2 0.080 252.39 1.00x10−3 0.08x100 8.35 3.65x103 1.22x107 1.57x1014 S2 CNG 0.017 129.78 0.80x100 10.43 7.95x101 2.08x106 7.47x1010 S3 LNG 2.695 26.59 4.50x108 10.43 2.24x101 3.29x108 5.93x109 S4 H2 0.077 5.34 6.00x10−3 0.08x100 8.35 9.76x101 1.96x106 1.88x1010 S5 CNG 0.650 2.75 0.80x100 10.43 8.45x101 1.32x107 1.40x1010 S6 LNG 0.013 0.56 4.50x108 10.43 3.00x10-3 2.65x105 1.78x101 The Mach number calculations confirmed that only one of the presented scenarios (S6) would be realized in the subsonic regime (M<1). All the other scenarios represent transonic frow regimes (M>1), especially scenario S1, where M=3,650. At the same time, all the release scenarios appeared to create strongly turbulent, momentum- dominated jests (Re>2000). The received Froud numbers confirm their downward trend in subsequent scenarios (around one in order of magnitude), the relevance of which is confirmed by the decreased turbulency. The presented results confirm that, together with the release velocity decreasing, all of the dimensionless numbers decrease also. 𝑀𝑀 = 𝑈𝑈 𝑎𝑎 (1) 𝑅𝑅𝑅𝑅 = 𝜌𝜌0𝑈𝑈𝑈𝑈 𝜇𝜇 (2) 𝐹𝐹𝐹𝐹 = 𝑈𝑈 𝑔𝑔𝑈𝑈 2 (3) 351 3. CFD simulations Based on the scenarios prescribed above, CFD simulations with the FDS 6 code were prepared. All the simulations were prepared in 3D model of the enclosure, using the software FDS, created by NIST (McGrattan et al., 2019). The scenarios assumed that the gases were released in a hall room of 10.6 m high and an area of 500 m2 (a volume of 5,300 m3). The hall security system provided for individual gas detectors and emergency ventilation with a capacity of 12 air volumes per hour. Ventilation started automatically, 40 s after the gas detection. The aim of the study was to analyse the conditions of gas propagation in the event of their emergency release in the hall space before starting emergency ventilation. The level of LEL condensation is presented in pictures in black to make it easier to understand them. A) B) Figure 3. Hydrogen release; A) 1 mm opening; B) 6 mm opening; t=40 s. A) B) Figure 4. CNG release; A) 1 mm opening; B) 6 mm opening; t=40 s. A) B) Figure 5. LNG release; A) 1 mm opening; B) 6 mm opening; t=40 s. After this time, the emergency ventilation was expected to be activated. Presented simulation results demonstrate the differences in the analysed H2, CNG, and LNG gases distribution in the enclosure, depending on the opening release diameter, within 40 s from the damaged bus installation. It is visible that in the case of 1 mm opening diameter, all the gases reached the ceiling, but their condensation doesn’t reach the LEL. In contrast, the 6 mm diameter opening causes a high exceedance of the LEL, mostly in the case of hydrogen release. These observations suggest that the orifice diameter could be a significantly important parameter in the explosive enclosures safety considerations in the event of the accidental release of flammable gases. The proposed solution of the identified hazard limitation could be the implementing of an obligatory limitation of the buses’ gas installations diameters to the indicated safe level. 4. Conclusions Hydrogen, CNG, and LNG gases were analysed under accidental release conditions. Six scenarios of bus gas fuels dispersion in a nominally closed space were undertaken. Mach, Reynolds, and Froud numbers calculation allowed for indicative regimes of the released jets. The results confirm a significant influence on the hydrogen dispersion release rate is down to the orifice diameter, and it has been proven that too large a gas outflow could create a dangerous situation, and that a gas concentration above LEL could appear before the emergency ventilation can activate. The suggested solution could be an implementation of an obligatory limitation of the 352 bus gas installation diameters. The observations presented in the article are important in practice for the design of protective measures against hydrogen explosions. Nomenclature D – outflow diameter (m), g - gravitational acceleration (9.81 m/s2), Mass – gas mass in the tank (kg), LEL – lower explosive limit (% vol, kg/m3), P – pressure in the tank (bar), U - velocity at the orifice (m/s), V - release outflow (kg/s), Vol – tank volume (m3), a - speed of sound (340.30 m/s), ρ0 – gas density (kg/m3), μ– dynamic gas viscosity (μPa·s). References Andrzejewski, T., 2019, Podstawy inżynierii chemicznej i procesowej. Gdańsk: Wydawnictwo Politechniki Gdańskiej. Brennan, S. L., Makarov, D. V. and Molkov, V., 2009, Safety assessment of unignited hydrogen discharge from onboard storage in garages with low levels of natural ventilation, International Journal of Hydrogen Energy. 22, Vol 3, 353–359. doi: 10.1016/j.jlp.2008.12.007. Brennan, S. and Molkov, V., 2013, Safety assessment of unignited hydrogen discharge from onboard storage in garages with low levels of natural ventilation, International Journal of Hydrogen Energy. 38, Vol 9, 8159– 8166. doi: 10.1016/j.ijhydene.2012.08.036. Brzezińska, D., 2019, LPG cars in a car park environment—how to make it safe, International Journal of Environmental Research and Public Health, 16, Vol 6. doi: 10.3390/ijerph16061062. Brzezińska, D. (2021) ‘Hydrogen dispersion phenomenon in nominally closed spaces’, International journal of hydrogen energy, 46, 28358–28365. doi: 10.1016/j.ijhydene.2021.06.061. Brzezińska, D., Dziubiński, M. and Markowski, A. S., 2017, Analyses of LPG Dispersion during Its Accidental Release in Enclosed Car Parks, Ecological Chemistry and Engineering S, 24, Vol 2, 249–261. doi: 10.1515/eces-2017-0017. Dorofeev, S.B., 2007, Evaluation of Safety Distances Related to unconfined hydrogen explosions, International Journal of Hydrogen Energy, 32, Vol 13, 1–12. doi: 10.1016/j.ijhydene.2007.04.003. Fischer, H. B., 2013, Mixing in inland and coastal waters. 1st, oryginal edn. Edited by Elsevier. doi: https://doi.org/10.1016/C2009-0-22051-4. Hajji, Y., Bouteraa, M., 2015, Natural ventilation of hydrogen during a leak in a residential garage’, Renewable and Sustainable Energy Reviews, 50, 810–818. doi: 10.1016/j.rser.2015.05.060. Hajji, Y., Jouini, B., 2015, Numerical study of hydrogen release accidents in a residential garage, International Journal of Hydrogen Energy. 40 Vol 31, 9747–9759. doi: 10.1016/j.ijhydene.2015.06.050. Kaplan, R. and Kopacz, M., 2020, Economic Conditions for Developing Hydrogen Production Based on Coal Gasification with Carbon Capture and Storage in Poland, Energies, 13, 1–20. doi: 10.3390/en13195074. Klassen, M. and Ph, D., 2020, Modern Vehicle Hazards in Parking Structures and Vehicle Carriers. Available at: https://www.nfpa.org/News-and-Research/Data-research-and-tools/Building-and-Life-Safety/Modern- Vehicle-Hazards-in-Parking-Garages-Vehicle-Carriers, accessed 24.05.2021. Krishnapisharody, K. and Irons, G. A., 2013, A critical review of the modified Froude number in Ladle Metallurgy, Metallurgical and Materials Transactions B: Process Metallurgy and Materials Processing Science, 44, Vol 6, 1486–1498. doi: 10.1007/s11663-013-9943-4. Lach, A. W., Gaathaug, A. V. and Vaagsaether, K., 2020, Pressure peaking phenomena: Unignited hydrogen releases in confined spaces – Large-scale experiments’, International Journal of Hydrogen Energy, 45 Vol 56, 32702–32712. doi: 10.1016/j.ijhydene.2020.08.221. Makarov, D., Hooker, P., 2018, Deflagrations of localised homogeneous and inhomogeneous hydrogen-air mixtures in enclosures, International Journal of Hydrogen Energy. Elsevier Ltd, 43, Vol 20, 848–9869. doi: 10.1016/j.ijhydene.2018.03.159. Makarov, D., Shentsov, V., 2018, Pressure peaking phenomenon: Model validation against unignited release and jet fire experiments, International Journal of Hydrogen Energy. 43, Vol 19, 9454–9469. doi: 10.1016/j.ijhydene.2018.03.162. McGrattan, K., 2019, Fire Dynamics Simulator User’s Guide, NIST Special Publication 1019 Sixth Edition, p. 347. Available at: https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication1019.pdf. accessed 24.05.2021. 353 Molkov, V.; Shentsov, V., 2016, Regimes of hydrogen jet fire in an enclosure with two vents, 6th Int. Symp. Non- Equilib. Process. Plasma Combust. Atmospheric Phenom. Sochi, Russia: Torus Press, 129–136. Molkov, V., 2012a, Fundamentals of Hydrogen Safety Engineering I. London, UK: Vladimir Molkov & bookboon.com. doi: 10.1016/B978-0-08-087872-0.00418-2. Molkov, V., 2012b, Fundamentals of Hydrogen Safety Engineering II. London, UK: Vladimir Molkov & bookboon.com. Molkov, V., 2014, Hydrogen non-premixed combustion in enclosure with one vent and sustained release: Numerical experiments, International Journal of Hydrogen Energy. 39, Vol 20, 0788–10801. doi: 10.1016/j.ijhydene.2014.05.007. Molkov, V. and Bragin, M., 2015, Hydrogen-Air Deflagrations: Vent Sizing Correlation For Low-Strength Equipement And Buildings, International Journal of Hydrogen Energy. 40, Vol 2, 1256–1266. doi: 10.1016/j.ijhydene.2014.11.067 Molkov, V. and Saffers, J. B., 2013, Hydrogen jet flames, International Journal of Hydrogen Energy. 38, Vol 19, 8141–8158. doi: 10.1016/j.ijhydene.2012.08.106. Molkov, V., Shentsov, V. and Quintiere, J., 2014, Passive ventilation of a sustained gaseous release in an enclosure with one vent, International Journal of Hydrogen Energy. 39, Vol 15, 8158–8168. doi: 10.1016/j.ijhydene.2014.03.069. Najjar, Y. S. H., 2013, Hydrogen safety: The road toward green technology, International Journal of Hydrogen Energy. 38, Vol 25, 10716–10728. doi: 10.1016/j.ijhydene.2013.05.126. Prasad, K., 2014, High-pressure release and dispersion of hydrogen in a partially enclosed compartment: Effect of natural and forced ventilation, International Journal of Hydrogen Energy. 39, Vol 12, pp. 6518–6532. doi: 10.1016/j.ijhydene.2014.01.189. Prasad, K., Pitts, W. M. and Yang, J. C., 2011, A numerical study of the release and dispersion of a buoyant gas in partially confined spaces, International Journal of Hydrogen Energy. 36, Vol 8, 5200–5210. doi: 10.1016/j.ijhydene.2011.01.118. Rodionov, A., Wilkening, H. and Moretto, P., 2011, Risk assessment of hydrogen explosion for private car with hydrogen-driven engine, International Journal of Hydrogen Energy. 36, Vol 3, 2398–2406. doi: 10.1016/j.ijhydene.2010.04.089. Saffers, J. B. and Molkov, V. V., 2014, Hydrogen safety engineering framework and elementary design safety tools, International Journal of Hydrogen Energy. 39, Vol 11, 6268–6285. doi: 10.1016/j.ijhydene.2013.06.060. Tolias, I. C., 2019, Best practice guidelines in numerical simulations and CFD benchmarking for hydrogen safety applications, International Journal of Hydrogen Energy. 44 Vol 17, 9050–9062. doi: 10.1016/j.ijhydene.2018.06.005. Weiner, S. C., 2014, Advancing the hydrogen safety knowledge base, International Journal of Hydrogen Energy, 39, 20357–20361. doi: 10.1016/j.ijhydene.2014.08.001. 354 lp-2022-abstract-048.pdf Enclosure Safety in a Case of Hydrogen, CNG, or LNG Accidental Release from a Bus Tank It is reported that the increasing usage of alternative fuels in vehicles (electricity, hydrogen, LNG, LPG etc.) has significantly altered vehicle fire characteristics (Klassen and Ph, 2020). For example, hydrogen ignition energy is ten times lower th... It was found that the release of gas fuel from the onboard storage tank could be a major risk, especially when initiated by an external fire. Such accidents could create different hazardous outcomes and are dependent on varying factors. The most dange... Therefore, theoretical verification of the gas dispersion phenomenon could be realized on the basis of three familiar dimensionless numbers analysis: Mach (M), Reynolds (Re), and Froud (Fr), which in many publications are used for gas plume characteri... The basic gases’ and bus tanks’ parameters are presented in Table 1. The main differences between the tanks are their internal pressure and the difference in the specific gravity of the gases causing significant differences in their capacity. The significant differences in velocity at the orifices appear due to the dif... Table 1. The gases’ and tanks’ parameters. Table 2. Reynolds, Mach, and Froud number values for the accidental releases. The Mach number calculations confirmed that only one of the presented scenarios (S6) would be realized in the subsonic regime (M<1). All the other scenarios represent transonic frow regimes (M>1), especially scenario S1, where M=3,650. At the same tim... Based on the scenarios prescribed above, CFD simulations with the FDS 6 code were prepared. All the simulations were prepared in 3D model of the enclosure, using the software FDS, created by NIST (McGrattan et al., 2019). The scenarios assumed that the gases were released in a hall room of 10.6 m high and an area of 500 m2 (a volume of 5,300 m3). The hall security system provided for individual gas detectors and emergency ventilation with a capacity of 12 air volumes p... Figure 3. Hydrogen release; A) 1 mm opening; B) 6 mm opening; t=40 s. Figure 4. CNG release; A) 1 mm opening; B) 6 mm opening; t=40 s. Figure 5. LNG release; A) 1 mm opening; B) 6 mm opening; t=40 s. After this time, the emergency ventilation was expected to be activated. Presented simulation results demonstrate the differences in the analysed H2, CNG, and LNG gases distribution in the enclosure, depending on the opening release diameter, within 4... Hydrogen, CNG, and LNG gases were analysed under accidental release conditions. Six scenarios of bus gas fuels dispersion in a nominally closed space were undertaken. Mach, Reynolds, and Froud numbers calculation allowed for indicative regimes of the ...