International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012)

Consequence Assessment of Vapour cloud Explosion Involving Hydrogen Release

A.R. Soman1, G.Sundararaj2
1

Research Scholar, Associate Professor, Department of Mechanical Engineering, PSG College of Technology, Coimbatore, Tamil Nadu, India The equivalent TNT mass model, TNT multi energy model and Baker-Strehlow models are used to estimate the overpressure resulting from VCE. We used Probit analysis to estimate the impact of VCE on people and structures [10, 11]. The probable consequences resulting from a gas fuel leak are shown in figure 1.

Abstract This work is concerned with the consequence analysis of vapour cloud explosion (VCE) due to accidental release of hydrogen (H2) from the hydrogen holder in a chloralkali plant. Hydrogen is highly flammable if released accidentally may lead to flash fire or VCE which has the potential to damage people, equipment and facilities in the vicinity of the hydrogen holder. The equivalent TNT, TNO multi energy and Baker-strehlow models are used to estimate the overpressure from the explosion. The probit equation is used to estimate the fatalities of people and damage of facilities from overpressure at different distance from the centre of explosion. The people working at a location of 25 m from the centre of explosion may get affected with different fatality range from 8 to100 %. The structure at a radial distance of 100m from the center of explosion may get minor damage with different probability from 32 to 100 %. The findings of the vulnerability analysis may be used to evaluate the improvements needs on the site and to ensure the safe design, position and location of existing and new structures. Keywords Hydrogen release, Explosion, Consequence analysis, Vulnerability analysis, Probit equation.

I. INTRODUCTION Vapour cloud explosions (VCE) are one of the most serious hazards in chemical process industries [1]. When a large quantity of flammable gas or vapor is accidentally released in to atmosphere it may form a vapour cloud and if its ignition is delayed (5-10 min) could produce a vapour cloud explosion. The damage effects of a vapour cloud explosion are mostly due to the overpressure that is created from the fast expansion of the combustion products. The overpressure is the most important causes of damage to people, equipment and facilities. Past accidents have revealed that, because of strong blast VCEs cause heavy damage to people, equipment and facilities [2-5]. The studies along with these lines have been described by several authors [6-9]. The factors influencing the evolution and intensity of an explosion are: (a) the type and the quantity of the flammable substance, (b) the time span from the onset of the leakage until the ignition (c) the configuration of the space where the leakage took place and (d) the position and the number of ignition sources in relation to the place of leak. 291
FIGURE 1 PROBABLE CONSEQUENCES RESULTING FROM A GAS FUEL LEAK

II. CONSEQUENCE ANALYSIS OF VCE WITH EXAMPLE CONDITIONS There are a lot of published articles about consequence analysis of vapour cloud explosions using mathematical models and computational fluid dynamics (CFD) modeling [12-16]. In this study, calculations for overpressures are performed for the accidental release of H2 gas from the holder having a capacity of 120 m3 maintained at above atmospheric pressure. The plant facility nearby hydrogen holder is shown in figure 2. In hydrochloric acid (HCL) synthesis unit, H2 and chlorine mixture are required to ignite inside the burner. Hydrogen blower is used to supply hydrogen gas from holder to HCL synthesis unit. A hypothetical scenario of a hydrogen release occuring from the hydrogen holder due to catastrophic failure of the holder is considered for this study.

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012) A scenario of flash back of flame mixture in the hydrogen line is assumed as ignition source. Flash back can results from failing to purge hydrogen line properly, using improper pressure, improper system operations etc .The worst-case scenario of VCE is assumed to be at the bottom of the holder (V: Figure 2) and it is considered as the centre of explosion. The first step in the assessment is to determine the overpressure released from the explosion. This energy is multiplied by a parameter (0 to 1) that accounts for the non-ideality of the explosion, and then divided with the heat of combustion of TNT, in order to calculate the equivalent TNT mass. The equivalent TNT mass is employed for the calculation of the shock wave in a specific distance from the source. A. The equivalent TNT mass (Kg) (1) Where G (kg) - mass of the hydrogen gas that takes part in the explosion (10.8Kg), c - heat of combustion of the hydrogen gas ( 120000kJ/kg ), - heat of combustion of (4,760 kJ/kg), and coefficient, arbitrarily taken as 0.1. B. Scaled distance The scaled distance (m/Kg1/3) (2) Where -equivalent TNT mass (27.22Kg) and distance from the center of explosion (25m). C.Overpressure released The overpressure of the shock wave (3) Where - the scaled distance (8.33 m /Kg1/3). At a location of 25 m from the centre of explosion the over pressure ( ) is found to be 0.126 bar. IV. TNO MULTI ENERGY METHOD The Netherlands Organization for Applied Scientific Research (TNO) has conducted extensive research in to blast models [18-21]. Multi-Energy method assumes that the vapor cloud explosion is composed of a number of subexplosions taking place inside specific areas of the cloud, corresponding to the various sources of blast that exist in the cloud [18]. The most important assumption of the method is that the strength of the explosion blast. The obstructed or partially obstructed regions (regions with high equipment density) will contribute to a high strength explosion blast. The remaining parts of the cloud will slowly burn, without a serious contribution to the strength of the blast [18, 21-22].The calculation of the over pressure is carried out using following algorithm. -

FIGURE 2 PLANT FACILITIES NEARBY HYDROGEN HOLDER

III. EQUIVALENT TNT MASS METHOD In this method, the power of the vapor cloud explosion equates to an equivalent mass of TNT (tri- nitrotoluene) that would produce the same explosive power [17]. First, the mass of the flammable gas in the cloud with concentrations between the lower and the upper flammability limits (LFL and UFL) is estimated. This mass is consequently multiplied by the heat of combustion to obtain the total available energy of combustion.

292

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012) A. Cloud dimensions The volume, V (m3) of the resulting vapor cloud (composed of flammable gas and air) is calculated from the reaction's stoichiometry and it is found to be 120*3=360m3(air is contains 21% O2). The radius of the resulting cloud, R (m), is derived from the volume of the cloud, being considered as a hemisphere and it is found to be 5.6 m. This radius determines the areas where the cloud will disperse. B. Obstructed region The approximate volume of the obstructed region (24*18*8.25 m) is found to be 3600 m3. Hence empty space available for the dispersion of the cloud is 3600 120 = 3480m3 which is larger than the volume of the cloud. Hence the entire cloud will be dispersed inside the obstructed region and assumed coefficient of an explosion blast as 10 (High strength-worst case). C. Energy of explosion The energy, (MJ) released by the explosion V. BAKER-STREHLOW METHOD (4) Where, c- the heat of combustion of hydrogen (120MJ/kg), V -the volume of the cloud(360m3), - the density of hydrogen (0.0899kg/m3) and the factor ( ) - the stoichiometry of the reaction (ratio of hydrogen : air = 1:3). The energy, released by the explosion is found to be 1300 MJ. D. The scaled distance The Sachs-scaled distance, (dimensionless) (5) Where, -distance from the center of explosion (25m), In the Baker-Strehlow method, the important parameter in the selection of the intensity of the explosion blast is the flame propagation speed. This is determined by (a) The way the flame front propagates, (b) The reactivity of the fuel, (c) the density of the obstacles, (d) the degree of confinement and (e) the energy-scaled distance from the gas blast center [23]. The procedures from the TNO multi energy methods were adopted for determination of vapour cloud dimensions and energy released from the explosion. The overpressure will be calculated as a function of the scaled distance, with the flame speed as a parameter. It is assumed that the flame expansion is 2-D and the flame speed is 1.77[24]. The scaled over pressure ( ) is calculated from the curve consist of scaled over pressure as a function of scaled distance [25] and it is found to be 0.3. At a distance of 25 m from the centre of explosion the blast over pressure is found to be 0.3 bar (Eq.6). VI. VULNERABILITY ESTIMATION The effects of an explosion on people and structures will be a function of the overpressure released from the explosion, as well as its positive phase duration. Table 1 shows the damage estimate for common structures based on overpressure [26]. Where, the Sachs-scaled over pressure (dimensionless) from the curve consist of scaled over pressure as a function of scaled distance is found to be 0.4[18] and - ambient pressure (0.1 MPa). The blast over pressure at a distance of 25 m from the centre of explosion is found to be 0.4 bars. F. Time duration of positive phase The time duration of the positive phase, (s) (7) Where, -Sachs-scaled positive phase duration (dimensionless) from the curve consists of positive phase duration as function of scaled distance is found to be 0.4 [18], Cs - the velocity of sound (340 m/s), - Energy released (1300MJ), -ambient pressure (0.1 MPa).The time duration of the positive phase, is found to be 0.0172 s.

-Energy released during the explosion (1300MJ), and - ambient pressure (0.1 MPa). The Sachs-scaled distance, is found to be 1.06.

E. Explosion overpressure The blast over pressure, = * (MPa) (6)

293

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012) Probit analysis provides a method to estimate the impact of explosion and different probit equations are shown in Table 2 [27].
TABLE 1 DAMAGE ESTIMATE FOR COMMON STRUCTURES BASED ON OVERPRESSURE

- Positive-phase duration (s) Whole-body displacement [11] Where,

Exposure level (bar) 0.0103 0.0276 0.09 0.276 0.689 20.68

Damage Typical pressure for glass breakage Limited minor structural damage Steel frame of clad building of slightly distorted Cladding of light industrial buildings ruptured Probable total destruction of building Limit of carter lip

Major structural damage

[12] Where +

Minor structural damage

[13] Where, +

The overpressure values obtained by the Multi-Energy Method are higher than those obtained by the Equivalent TNT method and Baker-Strehlow method. Also values produced by the Multi-Energy Method are nearer to the actual values observed based upon the damages that occurred from the explosions [28]. So Multi-Energy Method is considered for vulnerability calculations.
TABLE 2 PROBIT EQUATIONS [27]

VII. RESULTS AND DISCUSSIONS This study deals with the impact of hydrogen release from the hydrogen holder followed by vapour cloud explosion. It discusses consequences of explosion and action that can be taken to evaluate the estimated effects on peoples and structures. The equivalent TNT mass model, TNO multi energy model and Baker-strehlow model are used to calculate the overpressure from the explosion and at a distance of 25m from the centre of explosion which have been found to be 0.13 bar, 0.40 bar and 0.30 bar respectively. Accordingly the pressure were estimated at a distance of 100m from the centre of explosion are 0.025bar, 0.05 bar and 0.09 bar (Fig 3). Table 3 summarizes the variation of the overpressure, and time duration of positive phase with distance.

Lung damage and death

[8] Where,

-Total overpressure on person depending on orientation ( ) (Vertical orientation) - ambient pressure (pa) - impulse ( ) - Peak overpressure( Pa) - Positive-phase duration (s) - Body mass (Kg) Eardrum damage Head impact Where, [9] Where, - Peak overpressure (Pa) [10]

- Peak overpressure( Pa) - impulse ( )

FIGURE 3 OVERPRESSURE AS A FUNCTION OF DISTANCES

294

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012)
TABLE 3 VARIATION OF THE OVERPRESSURE AND TIME DURATION OF POSITIVE PHASE WITH DISTANCES.

TNT Method

TNO Method

Bakerstrehlow Method (bar) 2.5 0.7 0.4 0.3 0.17 0.10 0.09

For structural damage, the percentage of major structural damage is estimated (Eq. 12) as 100% at a location 12m from the centre of explosion whereas it is zero at a distance of 75m (Fig.5). For minor structural damage, the percentage of damage is estimated (Eq.13) as100% at a location 12m from the centre of explosion whereas it is 32% at a distance of 100m (Fig.5).

(m) 5 10 15 20 25 50 75 100

(m/Kg1/3) 1.66 3.33 5.00 6.66 8.33 16.6 25.0 33.33

(bar) 3.16 0.64 0.26 0.17 0.13 0.05 0.03 0.02

bar 0.212 0.420 0.638 0.850 1.060 2.120 3.190 4.250

(bar) 3.20 1.20 0.70 0.40 0.18 0.09 0.05

(ms) 12.4 13.3 15.2 17.2 24.2 27.6 31.6

For probability of death due to lung damage, the percentage fatality is estimated (Eq. 8) as 99% at a location of 8m from the centre of explosion whereas it is zero at a distance of 12m (Fig.4). For probability of eardrum rupture, the percentage fatality is estimated (Eq. 9) as 99.4% at a location of 8m from the centre of explosion whereas it is 8% at a distance of 25m (Fig.4). For probability of death due to head impact, the percentage fatality is estimated (Eq.10) as 100% at a location of 8m from the centre of explosion whereas it is zero at a distance of 12m (Fig.4). For probability of whole-body displacement impact, the percentage fatality is estimated (Eq.11) as 62% at a location of 8m from the centre of explosion whereas it is zero at a distance of 12m (Fig.4).

FIGURE 5 OVERPRESURE EFFECTS ON STRUCTURES

FIGURE 4 OVERPRESURE EFFECTS ON PEOPLE FIGURE 6 BLAST RADIUS AFTER EXPLOSION

295

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012) TNO correlation method estimated that buildings within 12m radius will be significantly damaged as shown by the outer ring in figure 6. The people working within 8m radius will be significantly affected with various impacts as shown by the inner ring in the figure 6. VIII. CONCLUSIONS AND RECOMMENDATIONS This study presents the results of consequence analysis of VCE carried out for the accidental release of hydrogen from the hydrogen holder having a capacity of 120m3 in a cholr-alkali plant. Overpressure values obtained by the TNO multi energy method are higher than other two methods. The maximum overpressure value of 5.5 bars at 8 m distance from the center of explosion was estimated by TNO multi energy model whereas overpressure values 1 bar, and 4.2 bars are estimated by equivalent TNT mass model and Baker-strehlow model respectively. The people working at a radial distance of 25m from the center of explosion may get affected with different fatality range from 8 to100 %. The structure at a radial distance of 100m from the center of explosion may get minor damage with different probability from 32 to 100 %. The operators room in HCL plant was located inside this blast range where the probability of the eardrum rupture is 88% (Fig 6). To enhance safety the operators room in the HCL plant can be shifted to a safer location or introduced protective measures. Protective measures in explosion include new venting devices with internal flame arrestors to prevent flame venting, and new design guidelines that render explosion venting more feasible in large buildings and process structures. Many new blast resistant barrier and blast mitigation wall products are also available for mitigation purpose. The recommendations of preventive measures are 1. Purging of hydrogen line should be performed with an inert gas such as nitrogen to avoid the formation of flammable mixtures. 2. Suitable flash back arrester/flame arrester should be provided in hydrogen line to shuts off gas flow in the event of flash back. 3. Gas and flame detectors should be installed in the vicinity of hydrogen holder. Because of hydrogen flame is invisible special detectors are required. 4. Adequate ventilation should be provided in all hydrogen systems to eliminate/minimize the potential hazards and formation of combustible mixtures. 5. Venting of hydrogen should be done according to standard and regulations. 6. Use an air monitor equipped to detect hydrogen on regular basis. 296 7. Barriers or safeguards should be provided to minimize risks and control failure. 8. Hazard placards are posted on hydrogen storage facilities. ACKNOWLEDGEMENT The authors would like to thank the management of the company where the study was conducted for their support and suggestions. NOMENCLATURE : Total energy released by the explosion, J : Fraction of energy : Scaled impulse,Pa1/2skg-1/3 G : Mass of the flammable gas, kg Pa : Ambient pressure, MPa S : Blast verpressure, MPa s : Sachs-Scaled overpressure, Pr : Probit value r' : Sachs-Scaled distance R : Radius of cloud, m tp : Positive phase duration of explosion in seconds tp : Sachs-scaled positive phase duration V : Volume of vapour cloud, m3 X : Distance from the centre of the explosion, m : Scaled distance, m/kg1/3 c : Heat of combustion of the flammable gas, kJ/kg : Heat of combustion of TNT, kJ/kg CFD : Computational Fluid Dynamics HCL : Hydrochloric acid LFL : Lower flammability limits MTNT: Equivalent TNT mass, Kg TNT : Tri- nitrotoluene TNO : The Netherlands Organization TCC : Travancore Cochin Chemicals UFL : Upper flammability limits VCE :Vapour Cloud Explosion REFERENCES
[1 ] M.J. Tang and Q.A.Baker, Process Safety Progress, 18 (3), pp. 253240 (1999). [2 ] V.C. Marshall, Consequence analysis of explosion hazards, in: K.V.Raghavana and G. Swaminathan, (Eds.), Hazard Assessment and Disaster Mitigation in Petroleum and Chemical Process Industries, Oxford & IBH, New Delhi, Calcutta, PP. 69-82 (1996). [3 ] L.A.Medard, Accidental Explosions, Wiley &Sons, New York (1985). [4 ] G.Wells, Major Hazards and their Management, IChemE, Great Britain (1997). [5 ] CCPS, Guide lines for evaluating the Characteristics of Vapour Cloud Explosions, Flash Fires, and BLEVEs, AIChE, New York (1994).

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 11, November 2012)
[6 ] F.P.Lees in: S.Mannan,(Ed.), Loss Prevention in the Process Industries- Hazard Identification, Assessment,andControl,vols.13,Elsevier/Butterworth - Heinemann, Oxford (2005). [7 ] D.A.Crowl, Understanding Explosions, Centre for Chemical Process Safety, AIChE, New York (2003). [8 ] J.L.Woodward, Estimating the flammable Mass of Vapour Cloud, CCPS/AIChE, New York(1998). [9 ] R.K. Eckhoff, Expolsion Hazards in Process Industries, Gulf Publishing Company, Austin, TX, (2005). [10 ] Enrique Gonzalez Ferradas, Fernando Diaz Alonso, Juan Francisco Sanchez Perez, Marta Doval Minarro, Agustin Minana Aznar, & Jose Ruiz Gimeno, Process Safety Progress 28(2)179-189(2009). [11 ] D.I. Finney, PROBIT Analysis, Camebridge University Press, London (1971). [12 ] R.W. Prugh, Process Safety Progress 18 (1), 5-16(1999). [13 ] F.Dias,E. Alonso, Gonzalez Ferradas, Sanchez Perez, J.F., Minana Anzar, A., Ruiz Gimeno, J.and J.Martinez Alonso, J. Hazardou. Materials A137, 734-741(2006). [14 ] J.Lobato,P.Canizers, M.Rodrigo,and Saez, L.Linares, C, Int.J. Hydrogen. Energy 31, 1780-1790(2006). [15 ] E.Salzano, and V. Cozzani, Rel. Eng. Syst. Saf. 90, 271-284(2005). [16 ] N.R.Popal,C.A. Catlin,B.J. Arntzen, R.P. Lindstedt, B.H.Hjertager, T. Solberg,O. Saeter and A.C.Vanderberg, J. Hazardous Materials,45,1-25(1996). [17 ] W.E.Baker, P.ACox, P.S.Westine, J.J. Kulesz, and R.A. Strehlow , Explosion Hazards and Evaluation, Elsevier, New York(1983). [18 ] A.C Van den Berg, J. Hazardous Materials 12, 1-10(1985). [19 ] A.C.Van den Berg, X.J.M.Van Wingerden, and The, H.G, Transactions of IChemE. Part B, Vol. 69, 139-148(1991). [20 ] W.P.M.Mercex,A.C. Van den Berg, and D.Van Leeuwen, Application of correlations to quantify the source strength of vapour cloud explosions in realistic situations: Final report for the project GAMES, TNO Reoprt, PMLI998-C53, Rijswijk, The Netherlands(1998). [21 ] W.P.M.Mercex,A.C. Van den Berg, C.J. Hayhurst, N.J. Robertson,and K.C. Moran, J. Hazard.Mater 71, 301-319(2000). [22 ] A.C.Van den Berg and A.Lannoy,J. Hazard.Mater 34, 151171(1993). [23 ] Q.A.Baker, M.J.Tang,E.A. Scheir,and G.J. Silva, Process Safety Progress 15,106-109(1996). [24 ] A.J.Pierorazio,J.K. Thomas, Q.A. Baker and D.E.Ketchum, Process Safety Progress 24, 59-65(2005). [25 ] Q.A.Baker,C.M. Doolittle,G.A Fitzgerald and M.J. Tang, Process Safety Progress 17, 297-301(1998). [26 ] V.J.Clancey, Diagonostic Features of Explosion Damage, 6th International Meeting on Forensic Science, Edinburg, Scotland (1972). [27 ] TNO 1992a, Methods for the determination of possible damage, CPR16E, Director General of Labour,The Netherlands,ISBN 905307-052-4(1992). [28 ] J.A.Marc and E.K. Konstantinos, Fires, Explosions, and Toxic gas Dispersions Effects calculation and Risk analysis, CRC Press, Taylor and Francis Group, Boca Raton,FL 33487-2742(2010).

297