Computational Fluid Dynamics (CFD) Modeling and Analysis of


Download Computational Fluid Dynamics (CFD) Modeling and Analysis of


Preview text

processes

Article
Computational Fluid Dynamics (CFD) Modeling and Analysis of Hydrocarbon Vapor Cloud Explosions (VCEs) in Amuay Refinery and Jaipur Plant Using FLACS
Zaman Sajid 1,2,* , Muhammad Kashif Khan 3,4, Alireza Rahnama 2, Farzan Sahari Moghaddam 2, Kirti Vardhan 2 and Reema Kalani 2

1 Department of Business Administration, University of the People, 225 S Lake Ave, Pasadena, CA 91101, USA 2 Department of Process Engineering, Faculty of Engineering & Applied Science, Memorial University of
Newfoundland, St. John’s, NL A1C 5S7, Canada; [email protected] (A.R.); [email protected] (F.S.M.); [email protected] (K.V.); [email protected] (R.K.) 3 School of Mechanical Engineering, Sungkyunkwan University, 2066 Seobu-Ro, Jangan-Gu Suwon, Gyeong Gi-Do, Seoul 16419, Korea; [email protected] 4 School of Chemical Engineering, Sungkyunkwan University, 2066 Seobu-Ro, Jangan-Gu Suwon, Gyeong Gi-Do, Seoul 16419, Korea * Correspondence: [email protected]; Tel.: +1-(0)709-765-8844

Citation: Sajid, Z.; Khan, M.K.; Rahnama, A.; Moghaddam, F.S.; Vardhan, K.; Kalani, R. Computational Fluid Dynamics (CFD) Modeling and Analysis of Hydrocarbon Vapor Cloud Explosions (VCEs) in Amuay Refinery and Jaipur Plant Using FLACS. Processes 2021, 9, 960. https://doi.org/10.3390/pr9060960
Academic Editor: Chi-Min Shu
Received: 14 April 2021 Accepted: 26 May 2021 Published: 28 May 2021

Abstract: Process safety helps prevent the unexpected and unplanned release of flammable and toxic chemicals, leading to poisonous gas clouds, fires, and explosions. Vapor cloud explosions (VCEs) are among the most severe hazards to humans and the environment in process facilities. Therefore, process safety demands to use best and reliable techniques to model VCEs in process industries and storage tanks of flammable chemicals. In this regard, the Computational Fluid Dynamics (CFD) models are more appropriate, as these models provide three-dimensional (3D) modeling of all sequences of events in an accident. In this study, CFD is used to model VCE in two industrial accidents: the Amuay refinery disaster (happened in 2012) and the Indian Oil Corporation’s (IOC) Jaipur terminal (2009). This work studies 3D CFD modeling of flammable cloud explosion in the real-time configuration for both accidents. FLACS (FLame ACceleration Simulator), a CFD software, is used to simulate the loss of hydrocarbon containment, cloud formation, and explosion in both industrial case studies. The ignition locations and grid sizes were varied to analyze their influence on explosion overpressure, temperature, vapor velocity, and fuel mass. This work also investigated the effect of geometry complexity on the explosion. Results showed that, as opposed to the coarse grid, the fine grid provides more precision in the analysis. The study also reveals an explosion overpressure of the order 4–15 bar (g) for the given case studies. This study’s results can help perform a qualitative and quantitative risk assessment of the Amuay refinery accident and Jaipur fire. The simulation of different scenarios can help develop and improve safety guidelines to mitigate similar accidents.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Keywords: process safety; Computational Fluid Dynamics (CFD); gas explosion; vapor cloud explosion (VCE) simulation; Amuay refinery; Jaipur fire

Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

1. Introduction
Fuels are a significant energy source for power generation and transportation. In general, fuels such as petrol, natural gas, diesel, and petroleum gas have high demand. The need for energy is pushing every country to develop refinement facilities for crude oil processing and refining. The proliferation of processing industries has increased the potential hazards to the safety of people and process facilities [1]. Today, process safety intermingles in the design, production, and maintenance of every processing facility. However, despite all safety guidelines and measures, accidents are happening. The magnitude of these accidents ranges from minor damage to equipment or personnel to catastrophic

Processes 2021, 9, 960. https://doi.org/10.3390/pr9060960

https://www.mdpi.com/journal/processes

Processes 2021, 9, 960

2 of 18
accidents, which cause losses of human lives and damage properties worth billions of dollars [2].
In the oil and gas process facility, the risk can be due to the rapid and unexpected release of flammable process chemicals, which can cause toxic clouds and explosions. Even when the leaks are small, they lead to pollution and adverse health effects. Although hazardous conditions vary from facility to facility, the magnitude of damage depends upon the quantity and the type of released material. Process facilities such as refineries and tank farms, which hold large amounts of fuels, have an enormous potential for destruction [3]. One of the most devastating and destructive events that can take place in an oil and gas facility is a vapor cloud explosion (VCE). The VCE happens due to the ignition of a flammable cloud. The cloud is a mixture of gas, vapor, aerosol, mist, and it forms due to the loss of a certain quantity of combustible gas or vaporizing liquid [4]. The flame speed in VCE is very high, which creates considerable overpressure. Consequently, the overpressure causes bulk damage to people, buildings, and property in the vicinity of the explosion.
Moreover, due to its substantial adverse impact, a VCE can initiate a domino effect. A VCE can trigger secondary, tertiary, or even more VCEs in a domino effect called the cascading effect of VCE [5]. A better understanding of the overpressure propagation in a VCE can help develop risk mitigation techniques for minimizing damages due to VCE. There are five conditions for VCE to occur. Those are: the release of flammable material; the ignition source’s location with high energy; the size of the vapor cloud formed; flow with turbulence condition for accelerating the flame front; and the cloud’s obstacles [6].
There had been numerous accidents due to VCE, such as the Buncefield accident [7], the Flixborough disaster [8], the Jaipur oil depot fire [9], and the Amuay refinery accident [10]. The scope of this study is the accidents that occurred in the Jaipur oil depot and the Amuay refinery. In this study, different parameters that affect VCE are examined. Considering VCEs accidents in two industries, an explosion model using computational fluid dynamics (CFD) is developed. The model considers the complex geometries of both industrial cases.
Literature Review
One of the accidents studied in this paper is the Amuay refinery accident, in 2012, near Punto Fijo, Venezuela. The other accident is the fire in the Jaipur oil depot in 2009 in Jaipur, India. There have been several studies conducted to model such accidents. Few researchers performed the numerical simulation of the Amuay refinery accident using CFD [11]. Their research developed simulation scenarios based on different atmospheric conditions and provided a comparison among them. The results of their study showed that wind speed played an essential role in analyzing cloud behavior. The research helped to understand dense gas dispersion in the Amuay refinery accident. In their study, the location of the ignition source was a nearby street.
In another study, the CFD model was developed to study cloud dispersion and the explosion of gas release in storage terminals of an oil and gas depot [12]. The study assessed the impact of different flow rates, duration of released gas, and surrounding conditions on vapor cloud dimensions, such as diameter, explosive strength, and cloud height. The study results revealed that obstacles in a storage facility of oil and gas affect the vapor cloud’s shape and size. In case 2 of the study [12], the control room, pump house, building, and piping system were simulated. The study showed that it took 1084 s for the cloud to cover the whole plant for 50 kg/s leakages of gasoline from a 1 m2 origin. The study also showed that the vapor cloud diameter in the Jaipur oil deport fire was between 800 m to 900 m. In another study, research was conducted to analyze the influence of volume blockage ratio (VBR) on the Jaipur fire explosion [13]. The CFD was used to construct the Jaipur fire accident, and the leakage of fuel vapor from a valve connected to the piping system was studied at different flow conditions. The study showed that under calm conditions, the cloud traveled 600–900 m, and it took a simulation time of 2–3 h.

Processes 2021, 9, 960

3 of 18

The studies presented in the literature review section helped propose better process

design facilities with more robust production and maintenance plans. However, such

studies have not considered the variations in fuel type (light versus heavy fuel), within

the site geometry in explosion modeling. Fuel type plays an essential role in defining the

intensity of the explosion. Moreover, previous studies consider limited accident scenarios.

The site geometry of the accident and the location of an ignition source are essential

aspects that need to be considered in computational-based studies. For example, the heat

flux distributions of fire and the flame behavior are highly dependent upon the ignition

source’s location, as highlighted in the literature [14]. Hence, an ignition source site plays

a vital role in assessing the overpressure in a VCE. Therefore, it should be included in

the CFD modeling. The present work demonstrates the applications of CFD simulation

to simulate Amuay refinery and Jaipur fire accidents. This study aims to predict various

characteristics of VCE in both industrial accidents. To the best of the authors’ investigations,

there has been no study conducted on the comparative behavior of VCEs in both accidents,

and hence is the objective of current work. This work also discusses various technical

aspects of both accidents to understand gas explosions better and investigates factors

that affect the CFD model. In these two case studies, the variations in ignition locations,

temperature, pressure, vapor velocity, and the mass of combustible fuels are also studied.

This work also identifies proper grid size for higher precision and captures the effect of gas

composition and geometry complexity using different ignition locations in two accidents.

A methodological framework of this study is shown in Figure 1. As shown in Figure 1, as a

preliminary step of the study, different scenarios are developed in step 1 for both accidents

under investigation. In step 2, input information is retrieved. This information includes

the meteorological conditions and details on the release of hydrocarbon for both accidents.

Step 3 is aimed to define the domain and geometry for the simulation of both accidents.

Computational grid and initial conditions are described in steps 4 and 5, respectively, while

in the last step, simulations are performed, and results are analyzed. Further details of the

Processes 2021, 9, x FOR PEER REVIEWmethodology are presented in Section 3 of this paper.

4 of 20

Figure 1. A glimpse of the methodological framework for this study.
Figure 1. A glimpse of the methodological framework for this study.
2. Background
2.1. Vapour Cloud Explosion (VCE)
Combustion of a large amount of flammable vapor cloud involving more than one tonne of combustible fuel causes an explosion. As a result of quick energy release, the explosion causes to have a magnificent overpressure. Consequently, more severe fatalities

Processes 2021, 9, 960

4 of 18

This paper consists of two parts. The first part provides a conceptual background of accidents in the Amuay refinery and Jaipur fire, and the second part describes the methodological framework of the study.

Processes 2021, 9, x FOR PEER REVIEW

2. Background 2.1. Vapour Cloud Explosion (VCE)
In general, the combustion of a large amount of flammable vapor cloud involving more than one tonne of combustible fuel causes an explosion. As a result of quick energy release, the explosion causes to have a magnificent overpressure. Consequently, more severe fatalities occur due to the minimal time to revert to protective action [15]. The release of combustible materials in the atmosphere results in the formation of a vapor cloud. If the propagating flammable elements find a source of ignition, it results in a VCE. The VCE demands analytical, computational (with CFD tool), and experimental investigations to mitigate future accidents. The most likely outcome of a flammable fluid discharge over some particular time is to form a flammable vapor cloud. The failure of an under-pressure5 of 20 vessel, pipe, or other facilities can lead to instant dispersion of flammable vapors. The general event tree for cloud vapor explosion is presented in Figure 2.

Release of hydrocarbon

Instant ignition
Yes

Delayed ignition

Explosion

Consequence Fire

No No

Release to atmosphere

No

Fire

Yes

Yes

VCE

Figure 2. An event tree of VCE. Figure 2. An event tree of VCE.
2.2. Detonation and Deflagration
2.2. Detonation and Deflagration
The overpressure produced due to the fuel mixture’s fast combustion spreads through
thTehfeueol vmeirxptureressausraeshporcokdwucaevde tdhautecatuosetshteo ifguneiltemthiextfuureel ’ms ixfatustre.coTmhebsuhsotcioknwasvpereads throaungdhfltahmeeffuroenl tmmixotvuerteogaestahesrhaotcskupwearsvoenitchvaetlcoacuitys.eDs utoe itgonthities cthaerfaucteelrmistiixc,tudreeto.nTahtieonshock wavisecaonndsifdlaermede mfrornet dmeostvruecttoivgeetahsecroamtpsaurpeedrtsoodneicflvagerloatciiotny. IDnudeetfloagthraisticohna, trhaecrteeriisstthice, detonastuiobnsoisniccovnesliodceitryed( to detonation transition (DDT). It has been revealed that the DDT phenomenon occurred
busitniotnheisAqmuuiacyklryefitnraenrysfaocrcmideedntianntdo Jaaidpuetrofinraet[i1o6n].eTxhpelroesaiorens. tTrohnegprehaesnoonms beenhoinndisthcealled defleaxgisrtaetniocne otof DdDetToninattihoensetrianndsuistitorinal(VDCDETs).. FItorhaexsabmeepnle,rethveeraelewdasthaaht itghhe flDaDmTe pspheeendomenon(colcocsuerrrtoedthienstpheeedAomf usoauynrde)finnebroythacaccidideennttsa.nUdnJdaeipr suurcfhirceir[c1u6m].sTtahnecrees,atrheersetrios naghirgehasons behpinodsstibhieliteyxtihstaetndceeflaogfrDatDioTn iwnotuhledsheaivnedbuesetnritarlanVsCitiEosn.eFdotroedxeatomnpatlieo,nth[1e7r]e. wToarsedauhciegthheflame speelidke(lcihloosoedrotof DthDeTs, preeseedarochfesrosuhnadve) ipnoibnotethd taocchiadveinngtsa. nUonbdsetarcsleu-cfrheecipractuhmanstdaanvcoeisd,itnhgere is a hicgohngpeosstisoibnilsiutychthasatredseidfleangtriaaltibounildwinogusldinhaavfierebreaedniattrioannsriatidoinuesd[1t0o]. dTehtoesneartieosnea[r1ch7]. To reduficneditnhges lairkeebliehinogodmoodf eDleDdTin, trheissesatrucdhyerussihnagvtheepgoeionmteedtrytoofhtahveiancgciadnenotsb.stacle-free path and avoiding congestion such as residential buildings in a fire radiation radius [10]. These research findings are being modeled in this study using the geometry of the accidents.

2.3. Affecting Parameters

Processes 2021, 9, 960

5 of 18
2.3. Affecting Parameters
Different parameters analyzed in this study are discussed below.
2.3.1. Geometry Complexity One of the primary conditions in which the subsonic transforms into supersonic flame
speed is the geometry complexity (obstacles and partial confinement in the flame path). The geometry of an accident scene can significantly affect the behavior of the explosion, and the gas explosion concentration can vigorously differ according to the flame’s velocity. Changes in boundary conditions can result in a remarkable change within the propagation of gas. The obstacle in the path of the explosion propagation has a considerable impact on increasing the gas velocity. The location of the obstruction profoundly influences the propagation of the flame. Research shows that a faster propagation of the flame is noticed when the obstacles are closer to the ignition source [18].
2.3.2. Ignition Source According to a study, some of the significant sources of accidents in process industries
are listed below [19]: • flames being exposed to the environment, such as flares; • burning substances; • solid surfaces have a high temperature; • electrical sparks beside similar other means of discharge; • confinement having an adiabatic situation.
The minimum ignition temperature of the fuel-air mixture is highly critical. Furthermore, the ratio for mixtures of hydrocarbon mixtures being investigated at their various temperatures of the mixture has a crucial role in the explosion modeling [20].
2.3.3. Wind Effect A study investigating the accidents due to the VCEs shows that the wind speed was
low in more than 70% of accidents [21]. A low rate of wind speed causes the leakage of fuel by gravity. The results advise that nil or low wind speed conditions can affect the entire explosion risk in many oil and gas plants. The weather conditions are occasionally rare than low wind velocity. However, they are associated with gigantic clouds with a much higher risk of ignition.
Moreover, wind conditions with low speed allow cloud formation through the low release rates of vapor. This condition has the potential to cause accidents as there is a high likelihood of continuous release. Hence, the modeling of the risk of VCE in petrochemical industries needs to include wind conditions. In a review study on VCEs, researchers categorized various industrial accidents based on wind conditions [21]. In their study, both accidents (Jaipur and Amuay) were classified as those which occurred in nil/low wind conditions. The study also identified vapor release rates of 67 kg/s and 3 kg/s for Amuay and Jaipur accidents, respectively. The wind data in their study was based on meteorological records and the shape of the cloud formed. A low wind speed was defined as the wind speed that spreads the vapor cloud to an equal extent and in all fuel source directions, and this definition is adopted in the current study.
2.4. The Explosion at Punto Fijo Refinery, Venezuela
An explosion occurred at Amuay (Punto Fijo) refinery, the second-largest refinery in the world, on 25 August 2012, around 01:11 a.m. (GMT) in the northwest of Venezuela. A plant layout of the Amuay refinery is shown in Figure 3. The shock wave released by the explosion affected the neighboring residential area. There was significant damage to people and property. More than 50 people lost their lives, over 100 people were seriously injured, several light injuries were reported, and more than 1600 houses were damaged. Residents of over 200 homes were evacuated to safer places. The damage in monetary terms was more than one billion U.S. dollars.

Processes 2021, 9, 960 Processes 2021, 9, x FOR PEER REVIEW

6 of 18 7 of 20

FigFuirgeu3re. 3A. AnnOOvveerrvviieeww aannddlalayyoouut ot fotfhethAemAumayuRaeyfiRneerfiynleorcyatelodcianteVdeninezuVeelnae, zFualeclóan, SFtaaltceó, nCaSrtiaruteb,aCnaarciirtyu.bAannaapcpitryo.x-An appimroaxtiemloacteatliooncaotifotnheodf itshaestderiszaosnteeriszohnigehilsighhitgehdliignhaterdedinboaxr. eRdepbrooxd.uRceedprwodithucpeedrmwiistshiopnefrrmomissKioirntifBrohmushKainrtiMBihshursah, an
Klaus-Dieter Wehrstedt, Holger Krebs, Fuel Processing Technology; published by Elsevier, 2014, with permission from MisEhlsraev, iKelra. us-Dieter Wehrstedt, Holger Krebs, Fuel Processing Technology; published by Elsevier, 2014, with permission from Elsevier.
Amuay incident, a layout shown in Figure 4, began with a leak of high vapor pressure substances such as alkenes, liquefied natural gas (LNG), or liquid butane (or propane) and formed a denser flammable vapor cloud in the atmosphere, as shown in Figure 4. The dispersed cloud stuck near the ground and created fog under low wind conditions. A study reported the spread of dispersed fuel to a radius of 0.6 km, due to low wind speed [10]. The accumulation of the dispersed vapor cloud proved to be catastrophic when the cloud caught an ignition source. Researchers believed that the ignition source was most probably not inside the Amuay refinery but passing by vehicle [22]. The overpressure generated, as the result of the explosion, was of high magnitude. It crushed steel fences and cars, blown the refinery structure, and bent the infrastructure and vehicles. Based on the media sources, various initiating events led to make this accident a catastrophic one. Those events were the negligence towards the regular maintenance and ignoring the leak and associated smells in the plant before the incident [10].
An installed component’s functional failure is the most frequent hazard in the petrochemical industries, including valves, pumps, and piping. These operational failures occur due to an improper maintenance schedule, corrosion, and inadequate safety knowledge practices. Some witnesses in the refinery reported seeing a liquefied gaseous fuel leakage [22]. Other studies reported initiating events as malfunctioning of the valve, pumping system, and piping connected to the olefin tanks [23]. In summary, it was believed that the Amuay incident occurred due to the failure of a valve in piping. Venezuela’s socialist government reported that the sabotage and manipulation of several bolts from a gas pump caused the rupture, which initiated a gas leakage in the flange of a pump. Subsequently, a quick alkene leak occurred due to a 7.4 cm opening in a pipeline [24].
Dispersion of high vapor pressure olefin fuel started to accumulate a combustible cloud in the atmosphere after a valve’s failure, as shown in Figure 5. This cloud of olefin Figure 4. A satellite map sotfuvcakpotor ctlhoeudgrfoourmnadtidonueintoAimtsuhayigrheefirnedreynisnitVyetnheaznuealair,.FIatlcfóonrmSteadte,wChairtieruibrraengaucliatyr,fwogaswith no reproduced with permissiponarftriocmulKairrtsimBheullsuhannleMssisihtrcao, nKtlaauins-eDdiestoemr We eahdrdstietdivt,eHs olilkgeerhKyrdebrso,gFeunelsPurloficdeses.inDguTeecthonsolol-w wind ogy; published by Elseviers,p2e0e1d4,,wthiteh dpiesrpmeisrssieodn ffruoeml sEtlasretveiedr.to accumulate and covered a radius of nearly 0.6 km [10]. A passing-by vehicle was recognized as the most likely source of ignition in the Amuay

Processes 2021, 9, 960

7 of 18

Figure 3. An Overview and layout of the Amuay Refinery located in Venezuela, Falcón State, Carirubana city. An approx-

imate location of the disaastcecrizdoenneti.sThhigehbliglahstetdgiennaerreadtebdoxa.nReopvreordpurceesdswuriteh ipnertmheissriaonngfreoomf K15ir0ti kBPhuasthoan20M0i0shkraP,a, which

Klaus-Dieter Wehrstedt, Holger Krebs, Fuel Processing Technology; published by Elsevier, 2014, with permission from

Elsevier.

proved to be destructive for property and human lives [10].

Processes 2021, 9, x FOR PEER REVIEW

8 of 20

An installed component’s functional failure is the most frequent hazard in the petrochemical industries, including valves, pumps, and piping. These operational failures occur due to an improper maintenance schedule, corrosion, and inadequate safety knowledge practices. Some witnesses in the refinery reported seeing a liquefied gaseous fuel leakage [22]. Other studies reported initiating events as malfunctioning the valve, pumping system, and piping connected to the olefin tanks [23]. In summary, it was believed that the Amuay incident occurred due to the failure of a valve in piping. Venezuela’s socialist government reported that the sabotage and manipulation of several bolts from a gas pump caused the rupture, which initiated a gas leakage in the flange of a pump. Subsequently, a quick alkene leak occurred due to a 7.4 cm opening in a pipeline [24].
Dispersion of high vapor pressure olefin fuel started accumulating a combustible cloud in the atmosphere after a valve’s failure, as shown in Figure 5. This Olefins cloud stuck to the ground due to its higher density than air. It formed white irregular fog with no particular smell unless it contained some additives like hydrogen sulfide. Due to slow wind speed, the dispersed fuel started to accumulate and covered a radius of nearly 0.6 FiguFriegu4.reA4.saAteslaltietellimteampaopf ovkfamvpaop[r1o0crl]co.luAodupdfaofsorsmrimnaagtti-iobonyn ivinneAhAimmcluueaawyyarreseffirineneceroyrgyinniinzVeVedneneazesuzteuhleae,lamF,aoFlcsaótlnclóiSknteaSltyeta,stCoeau, rCricrauerbiorafunibagacnniitatyi,ocwintyai,snwthase reproredpurcoedduwceidthwpitehrmpeirsmsioisnsiforAnomfmrouKmairyKtiairBctihciBudhsehunsathn.aTMnhMieshibsrhlaar,saKt, Klgaeluansue-sDr-aDiteieteedtrearWnWeoehvhrserstretpedrdtet,,sHHsuoorllgegeeirrnKKthrreeebbrssa,, FnFuugeeelloPPfrro1oc5cee0sssksiniPngagTtToeec2chh0nn0o0ol-lkoPgya;, ogy; published by Elsevier,w20h1i4c,hwpitrhovpeerdmtiossbioendferostmruEcltsievveiefro. r property and human lives [10]. published by Elsevier, 2014, with permission from Elsevier.

FiguFriegu5r.e(a5). (Taa)nTkanfikrefi,r(eb, )(bw) iwnidndpupsuhsihninggththeefifrireettooootthheerr ttaannkkss,,((cc))fifirereaafffefcetcintigngoloeflienfitnantakns,k(sd,)(sdi)tessitheoswhionwg tinhrgeethtraenekstanks subjseucbtejedctteodfitroefiarfetearfttehrethaeccaicdceidnet,nrte, prerpordoudcuecdedwwitihthppeermrmisisssioionnffrroomm KKiirrttii BBhhuusshhaannMMisishhrara, ,KKlaluaus-sD-DieiteetreWr Weherhsrtesdtetd, Ht, oHl-olger
ger Krebs. Fuel Processing Technology; published by Elsevier, 2014, with permission from Elsevier. Krebs. Fuel Processing Technology; published by Elsevier, 2014, with permission from Elsevier.
2.5. Explosion at Indian Oil Corporation’s (IOC) Jaipur Terminal, India
The accident shown in Figure 6 occurred on 29 October 2009, in a giant tank at the Indian Oil Corporation oil depot, in Jaipur India, that held 8000 kiloliters of petrol. The accident’s primary reason was gasoline leakage from a valve connected to the pipeline

Processes 2021, 9, 960

8 of 18

Processes 2021, 9, x FOR PEER REVIEW

9 of 20

2.5. Explosion at Indian Oil Corporation’s (IOC) Jaipur Terminal, India
The accident shown in Figure 6 occurred on 29 October 2009, in a giant tank at the firIendiniatnhOe isltCoroargpeortaatniokns.oTilhdisepacocti,dinenJat icpluairmIneddiatw, tehlavtehleilvdes8,0a0n0dkisloevlieterrasl opfepoeptlreolw. Terheeinjuarcecdidaenndt’slopsrtimmaorryertehaasnonRws.a2s.8gabsiollliionne l(eInakdaigaenfrruopmeeasv).alTvheecoinnvneesctiegdatointghetepaimpelfionuen[d25t]h. at thTehiegfinrietieornupsoteudricnemwualstipfllaemtaenpkrsoaonfdecleocnttrinicuaeldefqouriopvmereantwoerekfi[r2in6]g. Tuhpeoefxpalnosiinotnercnreaaltecdom-
buasmtioasnsievnegfiinreebianlltthheatnceoavrebryedartheaese.nTthire ifnirsetailnlathioenstaonradgsetatratnedksasfiprreeiandthtoe nveaarriobuysfuoetlher tatnaknsk,ss. eTthtienignvaellsttihgeat6in0g,0t0e0amkileosltiitmerasteodf tfhuaetltohnerefirwea. sTahne ufinrceocnatruosleleddclaetaaksatgroepfohric90dmamina.ge inEaigrhatdyi-uosneoftonnenaerslyo2f gkamso.lIitnwe masigohbtsheravveedvathpaotritzheedv. aTphoervcalpoourdcleonugduelfveednttuhaelleynetixrpelofadceidlity, inwcliuthdiannginthteenpsiltaynet,qutriuvcaklenotfftlooa3d8itnognnaerseao,f aTnNdTpeaxrptsloosifvtehse. Tphipeemlinaneatgeermmeinnat ld. eGciadseodline cotnocleent ttrhaetipoentrionlethuemapreraodwuacsts7b.8urpnetrctoenatvtooid1.a2npoethrceernpto, twenitthiainl atchceidgeanstodliunee tfolafimrme [a1b6i]l.ity raTnhgeel[e1a3k]i.ng gasoline created a vapor cloud that exploded and started a fire in the storage
tankBse.loTwhisisacthciedseunmt cmlaaimryedoftwtheelvseeqliuveens,caenodf seevveenrtasl [p2e7o]p; le were injured and lost more than Rs. 2.8 billion (Indian rupees). The investigating team found that the ignition source 1.wasloflsasmoefpprroiomfaerleycgtraicsaollienqeucipomnteanint morefinrti;ng up of an internal combustion engine in the 2.neaersbcyaapreeaosf.gTahsoelfiinre finrotmhessetcoornadgeartyanckosnstapirnemadentot;various other tanks, setting all the 3.60,0t0h0ekoilpoelirtaetrisnogf sfutaeflfownathseofirdrea.iTnhede fiurseecleasuss;ed catastrophic damage in a radius of nearly
4.2 kmla.cIkt wofatsoooblsetrovteadckthleattthheesvitaupaotrioclno;ud engulfed the entire facility, including the plant, 5.trucfklaowfflsoiandtihneg eanregai,naenedripnagrtdseosfigthne; pipeline terminal. Gasoline concentration in the area 6.wasn7o.8tepaemrceanvtatiola1b.2lepfeorrceonpte, rwatitihnignftrhoemgathsoelisniteeflaanmdminabcialisteyoraf negssee[n1t3i]a.l areas.

FigFuirgeu6re. S6a. tSealltietlelimteampaopf oJafiJpauipruprlapnlatnint dinicdaictiantgintghtehleocloactaiotinonofolfelaekakoroirgiigninanadndthtehetotpopvvieiwewoof fththeetatannkkss. .TThheeaarereaaoonnthe matpheismInadpiias, IRnadjiaas,tRhaanjasStthaatne,SJtaaitpeu, rJacipituyr. Rcietyp.roRdepurcoedduwceidthwpiethrmpiesrsmioinssiforonmfroKmirtKi iBrthiuBshhuasnhaMnisMhirsah,rJao,uJornuarnl aolf oLfoLsossPsrevenPtrieovneinntitohneiPn rtohceePssroIcnedssusIntrdieuss;trpieusb; lpisuhbeldishbeydEblyseEvliseerv, i2e0r,1280,1w8,itwhitpheprmerimssiisosniofnrofrmomElEslesveiveire.r.
The flammable cloud was formed because a massive quantity of gasoline escaped from containment and then vaporized. Various parameters contributed to the explosion: the explosive properties; weather conditions; and ignition source location. The energy emitted in the explosion created a blast wave that changed the nature of the blast. In a

Processes 2021, 9, 960

9 of 18
Below is the summary of the sequence of events [27];
1. loss of primary gasoline containment; 2. escape of gasoline from secondary containment; 3. the operating staff was ordained useless; 4. lack of tools to tackle the situation; 5. flaws in the engineering design; 6. no team available for operating from the site and in the case of essential areas.
The flammable cloud was formed because a massive quantity of gasoline escaped from containment and then vaporized. Various parameters contributed to the explosion: the explosive properties; weather conditions; and ignition source location. The energy emitted in the explosion created a blast wave that changed the nature of the blast. In a blast wave, the pressure increases instantaneously and drops very fast to return to atmospheric pressure. In this case, the fundamental reason was the complete absence of the specific site operating procedures and leakage mitigation equipment [28].
Computational Fluid Dynamics (CFD) techniques are being progressively applied to better understand fire and explosion behaviors [29]. In CFD, mathematics is applied through algorithms and numerical techniques to understand fluid flow problems with or without chemical reactions [30]. Although analytical methods provide great usefulness in studying various cloud explosions parameters, a CFD concept is broadly utilized for multiple applications such as VCE [29].
CFD allows the modeling of a detailed site geometry of the VCE facility. Moreover, CFD enables modification of wind conditions, as per case studies, in large-scale modeling experiments [31]. However, there is no experimental data that could help study the continuous release of flammable material in calm wind conditions in which vapor transport is caused by gravity. A few studies have focused on using CFD to analyze VCE at the Buncefield oil depot accident in low wind conditions [32,33]. The only quantitative data used in these studies was vapor front progress; however, the movement of the vapor front does not affect the cloud structure itself. Another challenge is to model CFD when there is a long-range transport of cloud.
Moreover, the selection of grid size in CFD affects the precision of the results. To fill this research gap, this paper models Amuay refinery and Jaipur fire accidents using CFD. This work uses the actual fuel compositions at the time of accidents. Both accidents are reconstructed in this simulation study, and analyses are performed.
3. Modeling Methodology
The CFD models are the most significant models to perform 3D (3 Dimensional) modeling and allow using the equations that govern fluid motion for a vast range of complex scenarios [12]. FLACS (FLame ACceleration Simulator), developed by GexCon AS (https://www.gexcon.com/, accessed on 9 May 2021), is a frequently utilized CFD tool in chemical and petroleum industries for explosion modeling in process safety and risk analysis. The FLACS implements and solves the compressible conservation equations (such as momentum, mass, species, and enthalpy) on a 3D Cartesian grid using finite volume [17]. Among CFD software for explosion modeling, FLACS is a proper research tool for investigating explosion details through various influencing factors in a gas explosion, such as gas cloud size and ignition position [34]. Moreover, FLACS can also help analyze other affecting parameters, including flow dispersion, shock waves, and the cloud’s dimensions. Due to such features, FLACS v10.4 was used to model and simulate the Amuay refinery accident and Jaipur fire in this study.
3.1. Input Data for Simulation
In this simulation study, 3D (3 Dimensional) models of the Amuay refinery and Jaipur fire accidents were built. The momentum and mass equations for both accidents and numerical models were adopted from the literature [35]. The general specification of the geometry within different x, y, and z directions was 25.6 m, 8 m, and 8 m, respectively. Two

Processes 2021, 9, 960

10 of 18

fuel compositions were studied in the simulation. The first fuel was flammable octane,

i.e., the fuel composition closer to that of in Jaipur oil depot fire, compared to flammable

natural gas composition (i.e., CH4 91.3%; C2H6 7%, C3H8 1.3%) as the second fuel, i.e.,

the fuel composition closer to that of in Amuay refinery explosion. Initial and boundary

conditions were defined as 2 m/s for initial wind speed. The stability in the atmospheric

conditions was achieved by setting relative turbulence to zero. This condition helped to

Processes 2021, 9, x FOR PEER REVIEW

11 of 20

achieve calm/zero wind conditions, as discussed earlier.

In both studies, simulations were performed with coarse and fine mesh in which the

grid numbers and the graphical presentation of the grids developed are shown in Table 1 aTnadblFeig1.uNreu7m,breerspofecgtriivdesliyn. different dimensions for both coarse and fine meshing.

Mesh

X

Y

Z

Table 1. The number of grids in different dimensions for both coarse and fine meshing.

Coarse grid numbers

28

12

8

FMineeshgrid numbers

X

112

Y 48

Z 32

Coarse grid numbers

28

12

8

FinTewgroiddniuffmerbeenrst grid sizing, 1i.1e2., coarse and fine gri4d8s, were used for simu32lation to as-

sess the results’ precision based on the grid sizing parameter.

FFigiguurree77. .GGeeoommeettrryywwitithhccooaarrsseeggrridid(l(elefft)t)aannddfifninggrridid(r(rigighht)t.).

TTwhoe deixffpelroesniot ngsridfosrizfuinegl,cio.em., pcooasritsieoannsdinfinveargiroiudss,awspeerectus,sesducfhorassimthuelarteiolenastoedaspseresstshuerree,stuelmtsp’ eprraetcuisrieo,nflbuaidsevdeoloncitthye, ganridd fsuizeilnlgevpealr, awmeerteeer.valuated. Besides, the ignition locatioTnheweaxspalolssoioansssefossrefdueilnctohmepsoimsituiolantsioinsvsairnioceusitawspaesctasn, osuthcehrapsrtohfeourenledalsyeidnpflrueesnsucirneg, tpemarpamereateurrefo, rflbuoidthvaeclocicditeyn, tasn, dasfudeislcluesvseeld, wpererevieovuaslluya. tTehde. Befefseicdteosf, gtheeomigentirtyiocnolmocpaletixointy woanstahlesoexapssloessioend winatshaensaimlyuzleadtiboynscosinnsciedietrwinagstawnootihgenritpioronfsoouunrdcleyloincflautieonncsi,nig.ep.,acraasme 1et(e1r8 fmor, b5omth, 5acmci)daenndtsc,aasse 2di(s1c3umss,e5dmpr, e5vmio)u. sFlLy.ACThSeheafsfeaclrteoafdygeboemenetvraylicdoamtepdleaxsiatyvoanluathbele etxopolofsoirodnewvealsoapninaglyCzeFdD bcyodceonfosridvearriinogustwgaosidginsiptieornsiosonucracseeslo[c3a6t,i3o7n] sa,nid.e,.,thcearseefo1r(e1, 8sumch, 5vmali,d5amtio)nanisdncoatsreep2e(a1t3edmi,n5thmis, 5stmud).y.FLACS has already been validated as a valuable tool for developing CFD code for various gas dispersion cases [36,37] and, therefore, such v4a.lRideastuiolntsiasnndotDriespceuastseidonin this study.
4. ResTuhletshaingdheDstisaccuhsiseivoend pressures, using both fuels, during the explosion with coarse and fine mesh calculations are summarized in Table 2.
The highest achieved pressures, using both fuels, during the explosion with coarse aTnadblfien2e. RmeesushltscoaflcsuimlautiloatniosnawreitshucmoamrsaeraizneddfiinneTmaebslhei2n.g.

Table 2. Results of simulation with coarse and fiHneigmheesshtinPgr.essure for (Bar (g)) Case Study

A Fine Mesh

A Coarse Mesh

Highest Pressure for (bar (g))

CasCea1sOe Sctaundey

15.00

8.00–8.50

Case2 Octane

5.2A5–F5i.n5e0 Mesh

A3.C0o0a–r3s.e25Mesh

CaseC1aNsea1tOurcatalngeas CaseC2aNsea2tOurcatalngeas
Case 1 Natural gas

7.50–8.1050.00 4.050.25–5.50
7.50–8.00

5.580.0–06–.80.050 1.730.0–02–.30.025
5.50–6.00

Case 2 Natural gas

4.00

1.70–2.00

Table 2 indicates that grid sizing is a critical parameter that significantly influences

the precision of the analysis. In Table 2, Case 1 and Case 2 represent the different locations

of the ignition source in modeling, as discussed in Section 3.1 of this paper. In Table 2,

“Case 1 Octane” and “Case 2 Octane” both indicate the use of octane as fuel, while “Case

1 Natural gas” and “Case 2 Natural gas” both show the use of natural gas as fuel in re-

Preparing to load PDF file. please wait...

0 of 0
100%
Computational Fluid Dynamics (CFD) Modeling and Analysis of