Flare Manual
Flare Manual
#.0
#.2
T$%es Of Flares
Ground Flares
(.0 +.0
Flare )$ste
In5ouse )oft0are *'aila!le For Design Of Flare )$ste )a %le Calculations Co %arati'e )tud$ Of Flare )$ste s Of Different Refineries Flare Load Reduction /ased On Interloc-s Flare Gas Reco'er$ )$ste s
Descri%tion
11.1 Introduction 11.2 )$ste 11.3 Co %. )election 7 )uction Control 11.# O%erating Feed!ac11.(*ttac5 ents
12.0
1.0
INTRODUCTION "ri ar$ function of Flare is to con'ert fla a!le: toxic or corrosi'e 'a%ours to less o!;ectiona!le co %onents !$ co !ustion. Dis%osal of co !usti!le gases: 'a%ours and li3uids !$ !urning is acco %lis5ed in flare. T5is ser'e as an e ergenc$ dis%osal s$ste to eli inate excess 5$drocar!ons co ing to it due to< a= Release fro safet$ 'al'es !ecause of an e3ui% ent failure or !ecause of a;or %lant e ergenc$ suc5 as a %o0er failure: cooling 0ater failure or a %lant fire. != Lea-ages fro safet$ 'al'es or control 'al'es. c= Dis%osal of fla a!le li3uid>'a%our %roducts or !$ %roducts 05ic5 can not !e ar-eted. "rocess flares are used %ri aril$ in t5e oil and %etroc5e ical industries fro initial %roduction t5roug5 trans%ortation: storage: refining and %rocessing. Flares are also used for ot5er %rocess a%%lications 05ere 5$drocar!on e issions ust !e controlled. * fe0 exa %les include< )e0age digestors: coal gasification and li3uefaction: roc-et engine testing and 5ea'$ 0ater %lants. T5e ain o!;ecti'e of t5e flare 5as !een to disc5arge fla a!le and 0aste gases at a safe location and !urn t5e in order to %rotect t5e en'iron ent fro %ollution and 5a?ards. Flaring: !eing a critical o%eration in an$ %lants its design ust !e go'erned on strict safet$ %rinci%les. T5e 'arious %ro!le s associated 0it5 flare are< a= Glare< T5e flare generates a considera!le a ount of glare %articularl$ during e ergencies. T5is causes incon'enience to t5e surrounding %o%ulation and also gi'es t5e i %ression of unnecessar$ 0aste of resources. != ) o-e< ) o-e fro flares causes a considera!le a ount of en'iron ental %ollution. T5e s o-e does occur in ost of t5e flare s$ste due to inco %lete co !ustion of gases. c= Noise< ,ig5 noise le'els are created !$ flare s$ste during %lant u%set conditions. T5is again causes incon'enience to t5e surrounding %o%ulation. "rolonged ex%osure to excessi'e noise a$ cause ental irritation: fatigue and e'en deafness. No0 so%5isticated design of flare ti%s 5a'e greatl$ reduced t5e noise %ollution. d= Radiation< During %ea- flare: t5e radiation le'el fro t5e fla e could !e 5ig5. T5is affects t5e surrounding 'egetation and 5a!itation. e= Ground le'el concentration @GLC= ,ig5 GLC of co !ustion %roducts or un!urnt 5$drocar!ons causes da age to 5u an 5ealt5: deterioration of aterials and structures or creation of an area 05ere %lantAani al s%ecies cannot sur'i'e. 1it5 t5e gro0ing consciousness of %eo%le in e'er$ countr$ regarding en'iron ental %rotection: t5ere is a lot of %ressure to co e u% 0it5 !etter flare s$ste s to eli inate t5e %ro!le s entioned a!o'e. *TBO)",&RIC C&NTING< *not5er 0idel$ used dis%osal et5od is 'enting or direct disc5arge to at os%5ere. T5ere is so e difference of o%inion on t5e 3uestion of 05at a$ and 05at a$ not !e 'ented to at os%5ere. It is usuall$ considered safe to 'ent 5$drocar!on 'a%ours or gases lig5ter t5an air: exce%t 5$drocar!on 'a%ours or gases lig5ter t5an air: exce%t 5$drogen. )o e designers reco end 'enting of 'a%ours e'en as 5ea'$ as %ro%ane: %ro'ided it is 'ented fro an ele'ated 'ent stac- at a 5ig5 'elocit$. T5e 'ent stac- 5eig5t is selected so t5at t5e
concentration of 'a%our at grade is 0ell !elo0 t5e lo0er fla a!le li it of t5e 'a%ours. T5e relia!ilit$ for esti ation of ground le'el concentrations is 3uestiona!le and t5erefore flaring is t5e onl$ satisfactor$ et5od for safe dis%osal of 5$drocar!ons 5ea'ier t5an air. 2..0 CAUSES OF FLARE LOAD: "ressure Cessels: ,eat &xc5anges: o%erating e3ui% ents and %i%ing are designed to contain t5e s$ste %ressure. T5e design is !ased on t5e nor al o%erating %ressure at o%erating te %erature: t5e effect of an$ co !ination of ec5anical loadings t5at is li-el$ to occur: and t5e differential !et0een t5e o%erating and set %ressures of t5e %ressure relie'ing de'ice. O'er%ressure results !ecause of an un!alance or disru%tion in nor al flo0s of aterial and energ$ t5at causes accu ulation or loss in certain %art of t5e s$ste . T5e %rocess s$ste designer ust define t5e relief re3uired to %re'ent t5e %ressure in an$ %iece of e3ui% ent fro exceeding t5e axi u allo0a!le accu ulated %ressure. T5e causes of o'er %ressures are listed !elo0< Closed outlet on Cessels T5e inad'ertent closure of a !loc- 'al'e on t5e outlet of a "ressure Cessel 05ile t5e %lant is on strea a$ ex%ose t5e 'essel to a %ressure t5at exceeds t5e axi u %er issi!le 0or-ing %ressure. If closure of 'al'e results in O'er%ressure: a "ressure Relief 'al'e s$ste is re3uired. Inad'ertent Cal'e O%ening< In ad'ertent o%ening of an$ 'al'e fro 5ig5 %ressure s$ste to lo0 %ressure s$ste a$ cause an O'er%ressure in connecting lines or recei'ing s$ste . * %ressure relief 'al'e is re3uired to %rotect t5e s$ste . C5ec- 'al'e alfunction< T5e failure of c5ec- 'al'e to close t5e re'erse flo0 ust !e considered 05ere %otential of !ac- flo0 of 5ig5 %ressure fluid exists. 8tilit$ Failure< Loss of an$ utilit$ ser'ice: 05et5er unit0ide or local a$ cause o'er%ressure. * %otential failure a$ also lead an e3ui% ent to o'er%ressure. In ost of t5e case t5e utilit$ failure go'erns t5e axi u failure load. "ossi!le utilit$ failure 7 e3ui% ent affected are listed !elo0< Utility Failure &lectric Cooling 1ater Inst. *ir )tea Fuel @Oil: Gas= Inert Gas o Equipment Affe te! "u %s for circulating cooling 0ater: !oiler feed 0ater: 3uenc5 or reflux. Fans for air cooled exc5anges: Cooling to0ers or co !ustion air. Condensers for "rocess or 8tilities )er'ice coolers for "rocess fluids: lu!ricating oil or seal oil. Trans itters and Controllers: "rocess regulating 'al'es: *lar and s5utdo0n s$ste Tur!ine dri'ers for Rotating &3ui% ent: Re!oilers: Reci%rocating %u %s: &;ectors /oilers for "rocess stea : Re!oilers: Gas Tur!ines: Co %ressors )eal: Catal$tic reactors
o o
"artial Failure< "artial %reotectice credit can !e ta-en for nor all$ o%erating %arallel e3ui% ents or stand
o o o
o o o
!$ ser'ices t5at 5a'e t0o unrelated source of energ$ to t5e dri'ers. )tand!$ ser'ices cannot !e considered totall$ relia!le 5ence it does not suffice insurance against o'er%ressure. &lectrical or Bec5anical Failure< T5e failure of electrical or ec5anical e3ui% ent t5at %ro'ides cooling or condensation in %rocess strea s can cause o'er%ressure in %rocess 'essels. Reflux Failure< Reflux failure causes condensers flooding or loss of coolant resulting in an o'er%ressure. ,eat exc5anger tu!e failure< 15en tu!e side %ressure of an exc5anger is uc5 lo0er t5an t5e s5ell side: ru%ture of tu!e 0ill o'er%ressure t5e tu!e side. T5er al s5oc-: 'i!ration: corrosion a$ also cause tu!e failure. T5is ust !e %rotected 0it5 a %ressure relief 'al'e.: C5e ical reaction< O'er %ressure a$ occur !ecause of un!alance reaction in a reactor. )o%5isticate safet$ 'al'es are used in t5ese cases. ,$draulic &x%ansion< T5er al relief 'al'es are %ro'ided to o'erco e o'er%ressure !ecause of rise in te %erature. "lant Fires< *n$ %rocess e3ui% ent in an o%erating %lant t5at 5andles or %rocesses fluid a$ get ex%osed to fire. Continuous 'a%our generation 0ill lead to o'er %ressure in s$ste . "ressure relief 'al'e is re3uired to %rotect t5e 'essels releasing t5e generated 'a%or. Local "o0er Failure< T5ere ig5t !e local %o0er failures in unitsAe3ui% ents 05ic5 0ill cause distur!ances to local e3ui% ents. T5is a$ result in o'er %ressure in s$ste . DESI#N #UIDELINES T5e %er issi!le li its on %ollution: radiation and ot5er effects on en'iron ent 5a'e !een laid do0n considering 'arious as%ects. *lt5oug5 t5ere is no definite 3uantitati'e li it on flare and s o-e e ission fro flare: ost aut5orities de and t5at t5ese s5ould ini u . For t5er al radiation: noise and ground le'el concentrations follo0ing guidelines s5ould !e follo0ed< Radiation< Baxi u %er issi!le le'el is a function of t5e lengt5 of ex%osure: factors in'ol'ing reaction ti e and 5u an o!ilit$. Follo0ing li its of 5eat intensit$ radiations s5ould !e o!ser'ed< "er issi!le Design Le'el @.= Conditions .1A 2 /tuA,r> ft2 1.(4 (00 Continuous ex%osure 0it5 a%%ro%riate clot5ing #.23 1(00 & ergenc$ action lasting for se'eral inutes for %ersonnel 0it5 %ro%er clot5ing !ut 0it5out s5ielding. +.31 2000 & ergenc$ action lasting u%to one inute 0it5out s5ielding !ut a%%ro%riate clot5ing. 6.#+ 3000 Radiation le'el at an$ location 05ere %eo%le 5a'e access. For exa %le at grade !elo0 t5e flare or a ser'ice %latfor s near!$ to0er= ex%osure s5ould !e li ited to a fe0 seconds sufficient for esca%e onl$.
".0
!e 3.1
Radiation le'el on structure and in areas 05ere o%erators are not li-el$ to !e %erfor ing duties and 05ere s5elter fro radiating 5eat is a'aila!le @for exa %le !e5ind e3ui% ent=. Note< On to0ers or ot5er ele'ated structures 05ere ra%id esca%e is not %ossi!le: ladders ust !e %ro'ided on side a0a$ fro flare 05en . is greater t5an +.31 .1A 2.@Reference *"I R" > (21: Fift5 edition=. T5e 5eat radiation le'el for %ersonnel are 0it5 account ta-en of reaction ti e and 5u an o!ilit$ ti e of !et0een 4 and 1( seconds fro t5e onset of 5eat radiation. In e ergenc$ release: a reaction ti e 3 > ( seconds a$ !e assu ed and ( > 10 seconds 0ould ela%sed !efore indi'idual could see- co'er or de%ort fro t5e area: 05ic5 0ould result ex%osure ti e 4 > 1( seconds. &x%osure ti e necessar$ to reac5 t5e "ain t5res5old. Ra!iati$n inten%ity Time t$ pain t&re%&$l! /T8A5r>ft2 .1A 2 )econds ((0 1.2# +0 2#0 2.33 #0 620 2.60 30 1(00 #.23 1+ 2200 +.6# 6 3000 6.#+ + #200 11.+2 # +300 16.42 2 T5e radiation le'els ta-e into account t5e influence of clot5ing: !ut excludes t5e contri!ution of 0ind c5ill factor: a!ilit$ of %ersonnel to face a0a$ fro t5e radiation source and lac- of large o%en areas on offs5ore %latfor s. T5e 5eat radiation figures are exclusi'e of an allo0ance for solar radiation. In our countr$ solar radiation 0ould account for a!out 3(0 /tuA,r ft2 5eat and it is reco ended t5at for continuous ex%osure solar radiation s5ould !e ta-en into consideration. In t5at case allo0a!le 5eat radiation !$ co !ustion s5ould !e ta-en as (00 > 3(0 D 1(0 /tuA,r > ft2. For ot5er conditions for calculations solar radiation s5ould not !e su!tracted. 3.2 N$i%e: T5e ex%osure 1 inutes for noise %rescri!ed in so e of t5e 0ell acce%ted standards are gi'en !elo0< Dail$ ex%osureO),* Noise Regulations NIO), Noise Regulations ,rs d/a d/a 4 60 4( # 6( 60 2 100 6( 1 10( 100 0.( 110 10( 0.2( 11( 110
1(.22
(000
3.3
#LC $f t$'i
T5res5old li its of GLC for certain toxic su!stances @gases and 'a%ors= are as follo0s< #a% $r )ap$r **+ #a% $r )ap$r **+ *cetalde5$de 200 *cetic acid 10 *cetic an5$dride ( *cetone 1:000 *crolein 0.( *cr$lonitire 20 * onia 100 * $l acetate 200 * $l alco5ol 100 *niline ( *rsinic 0.00( /en?ene 3( /en?$l c5loride 1 /ro ide 1 /utadiene 1000 /ut$l alco5ol 100 /ut$la ine ( Car!on dioxide (000 Car!on disulfide 20 Car!on onoxide 100 Car!on tetrac5loride 2( C5lorine 1 C5loro!en?ene 2( C5lorofor 100 Cresol @all iso ers= ( C$clo5exane #00 C$clo5exanol 100 C$clo5exanone 100 #a% $r )ap$r C$clo5exene Diacetone alco5ol 1:1 Dic5loroet5ane Diiso!ut$l -etone Di et5$lsulfate &t5$l acetate &t5$la ine &t5$l !ro ide &t5$l et5er &t5$lenedra ine &t5$lene dic5loride Fluorine Gasoline ,$drogen selenide Isodine Iso%ro%$la ine Bet5$l acetate Bet5$l alco5ol 200 2>Bet5$lox$et5anol Bet5$l$clo5exane Bet5$lc$clo5exanone Bet5$l a $l alco5ol Na%5t5a @coal or tar= %Nitroaniline Nitro!en?ene Nitrogen dioxide Nitrotoluene O?one #a% $r )ap$r **+ #00 (0 100 (0 1 #00 2( 200 #00 10 100 0.1 (00 0.0( 0.1 ( 200 2( (00 100 2( 200 1 1 ( ( 0.1 **+ #a% $r )ap$r C$clo%ro%ane O>Dic5loro!en?ene Diet5$la ine Di et5$laniline Diet5$lene dioxide &t5$l alco5ole @et5anol= &t5$l!en?ene <$l c5loride &t5$lene c5loro5$drin &t5$lene di!ro ide &t5$lene oxide 100 For alde5$de ,$dra?ine ,$drogen sulfide Iso%5orene Besit$l oxide Bet5$l acet$lene Bet5$l !ro ide 20 Bet5$l c5loride Bet5$lc$clo5exanol Bet5$l for ate Bet5$lene c5loride Na%5t5a @%etroleu = Nic-el car!on$l Nitroet5ane Nitro et5ane Octane "entane #a% $r )ap$r **+ #00 (0 2( ( 100 1000 200 1000 ( 2( ( 1 20 2( (0 1000 100 100 100 (00 (00 0.001 100 100 (00 1000 **+
"ro%$l -etone "5en$l5$dra?ine "5os%5ine "ro%$l acetate "ro%$l et5er "$ridine )t!ine )ulfur dioxide )ulfur onoc5loride 1:1: 2:2>Tetrac5loroet5ane Toluene @toluol= Tric5loroet5$lene #.0
"5enol "5osgene @car!on$l c5loride= "5or%5orous tric5loride "ro%$l alco5ol "ro%$lene dic5loride Euinone )t$rene )ulfur 5exafluoride )ulfur %entafluoride Tetranitro et5ane o>Toluidine
#.1
T,*E OF FLARES T5e ideal flare is a de'ice t5at !urns 5$drocar!ons co %letel$. T5e co !ustion s5ould !e s o-eless: least nois$ and 0it5 ini u incon'enience to co unit$ in ter s of radiation and lu inosit$ as%ects of flare. /ased on s%ecific re3uire ents 'arious flaring conce%ts 5a'e e erged and are !eing %ractised. Carious t$%es of flare are< Ele-ate! flare%: Con'entional %i%e flare< Con'entional flare consists of a %i%e stac- 0it5 a !urner on to%. "ilot !urners re ain lit near t5e !urner: 05ic5 ensure !urning of 5$drocar!ons fro t5e flare stac- during u%set conditions. )o e of t5e co on features of con'entional %i%e flares are %ro!le li-e s o-e: fla e lift off and 5ig5 %urge re3uire ents. T5is t$%e of flare s$ste s are si %le: %ro'en and suita!le for 'ar$ing loads. T5e noise le'el is also 'er$ less due to lo0 exit 'elocities. Coanda &ffect Flares< )-in>ad5esion effect: disco'ered !$ ,enri Coanda: is used in t5is t$%e of flares. T5is effect introduces !ot5 air and tur!ulence in t5e fla e area: 05ic5 ensures co %lete co !ustion. ,ig5 %ressure gas e;ected fro a narro0 slot follo0s a %rofile of cur'ed surface. T5e Coanda %rofile aids in for ation of a 5ollo0 c$linder of a gas and entrains air u%to t0ent$ ti es t5e gas 'olu e. T5e ain feature of t5is t$%e of flare s$ste s are s5orter fla e lengt5: sta!le: fla e: s o-eless co !ustion and gi'es lo0er 5eat e issi'it$. T5e ot5er ad'antage is t5at t5ese can 5andle s all 3uantities of li3uid dro%lets. ,o0e'er: t5is re3uires 5ig5 %ressure gas: 5a'ing lo0 turndo0n ca%a!ilities: gi'es 5ig5er noise le'el and re3uires 5ig5er aintenance. Fet Bix Flares< T5ese are a'aila!le in 'arious configurations. Bulti%le no??les 0it5 se'eral 5oles: 0ings 0it5 slotted o%enings or dou!le angle drilled no??les %roduce s0irling action to en5ance air entrain ent. Instead of air going to t5e gas: t5e flared gas is s%read out into a !ient air at 5ig5 'elocit$: %ro oting ra%id: tur!ulent ixing 0it5 air and t5us ensuring s o-eless co !ustion. T5e ad'antages of t5is t$%e of flares are: gi'ing s5ort: sta!le: s o-eless fla e and lo0 e issi'it$. T5ese are %ro'ided 0it5 0ind s5ielding and 5ence are less affected 0it5 0ind
'elocit$ and direction. ,o0e'er: t5e %ro!le s associated are of 5a'ing co %licated configuration: 5a'ing li ited turndo0n ca%a!ilities: noisil$ co !ustion and 5ig5er %ressure dro%s. *ir>*ssisted Flares< T5ese flares incor%orating s%ecial ixing 5eads t5at increase gas to air interface. *ir is !lo0n into t5e central duct allo0ing unrestricted %assage for t5e lo0 %ressure relief gas flo0ing u% t5e annular duct. T5e air is !lo0n 0it5 t5e 5el% of a axial fan. T5e onl$ utilit$ re3uired are %o0er for t5e fan and fuel gas for t5e %ilot !urner. *ir assisted flares can 5andle 5ea'ier gases: gi'es lo0 radiation and glare and 5a'ing longer ti% life. T5e disad'antages of t5is t$%e of flare are: it re3uires utilit$ source: 5a'ing 5ig5 o%erating cost and co %lex controls are re3uired for efficient o%eration.Inclusion of a rotating e3ui% ent to su%%l$ co !ustion air reduces t5e s$ste relia!ilit$. )tea *ssisted Flares< T5ese flares ac5ie'e t5eir s o-eless ca%a!ilities !$ ixing stea 0it5 t5e flare gases. T5e stea is introduced at one or se'eral %oints to induce secondar$ air and tur!ulence to i %ro'e ixing. In addition to rendering t5e fla e s o-eless: stea reduces t5e fla e te %erature: t5us reducing t5er al radiation. )tea assisted flares also can 5andle 5ea'ier gases and 5a'e longer ti% life. T5e stea re3uire ent 0ould de%end on flare gas olecular 0eig5t. T5e o%erating costs are 5ig5er due to stea consu %tion. Bulti>!urner flares< T5e gas strea is s%lit and !urnt in ulti%le !urners installed on co on stac-. T5is reduces fla e lengt5 and t5ic-ness 05ile increasing t5e fla e surface and 5el%s air or stea to %enetrate t5e fla e easil$: ensuring efficient and i %ro'ed fla e c5aracteristics. #.2 #r$un! Flare%: In ground flares: co !ustion occurs at grade: t$%icall$ inside refractor$ lined steel enclosures of circular or rectangular section. Flaring load is staged and eac5 stage co %rises of ulti%le ;ets. Due to 5ig5 cost and ca%acit$ li itations: t5ese are often used in con;unction 0it5 ele'ated flares. T5ere are essentiall$ designed to dis%ose off s all continuous loads 0it5 s o-eless: in'isi!le and less nois$ co !ustion. Ground flare is !asicall$ installed for< > Concealing co !ustion of continuous load > ) all load fro cost considerations %referred. T5e use of ground flare re3uires a clear distance of a!out 1(0 B fro an$ %rocessing unit or storage facilit$. T5e flo0 rates 05ic5 can !e 5andled !$ a ground flare are generall$ constructed of fire !ric- or of car!on steel lines 0it5 refractor$ aterial. FLARE S,STE+ CO+*ONENTS AND CONCE*TS: T5e essential ele ents of an$ flare s$ste are< o Flare 5eader o ..O. Dru o 1ater seal dru o Flare gas seal o Flare ti% o Ignitor T5e !asic sc5e e 0it5 co %onents is s5o0n in s-etc5 (.1. T5e flare 5eader collects
(.0
t5e aterial relie'ed !$ safet$ 'al'es for safe discuarge to t5e flare.T5e axi u allo0a!le !ac- %ressure 0ould !e< > 10G of t5e set %ressure for con'entional safet$ 'al'es > (0G of t5e set %ressure for !alance !ello0 t$%e safet$ 'al'es > 20G of t5e set %ressure for %ilot o%erated safet$ 'al'es > @Baxi u allo0a!le !ac- %ressure and set %ressure are ex%ressed as gauge %ressures=. T5e nu !er of flare 5eaders re3uired is decided !ased on te %erature: allo0a!le %ressure dro%: %ro%erties of 5$drocar!on: oisture content etc. If t5ere is a %ossi!ilit$ of li3uid disc5arge into t5e flare s$ste a ..O. Dru is %ro'ided i ediatel$ ad;acent to %rocess unit. For %rocess %lant 5andling lig5t 5$drocar!ons li-e C1: C2: C3: C# t5e li3uid !lo0 do0n is also sent to Flare I.)./.L. -noc- out dru . * suita!le Ca%ori?er *r stong t$%e or indirect et5anol is used for 'a%ori?ing t5e li3uid 5$drocar!on. In case: 'er$ lo0 te %erature gases fro a %articular unit is %ut to a co on flare 5eader in offsites due to econo ic reasons: a flare gas su%er5eater on t5e lo0 te %erature 5eader a$ !e re3uired to 5eat t5e gases. In addition to t5is a ..O. Dru is needed near t5e flare stac- to tra% an$ li3uid condensed en route to t5e flare. T5is is %ro'ided to %re'ent 5a?ards associated 0it5 !urning li3uid dro%lets esca%ing fro t5e flare stac-. Considera!le 3uantities of li3uids can accu ulate in flare s$ste s !ecause of t5e de0 %oint c5aracteristics of t5e co %lex ixture of 'a%ours disc5arged at 5ig5 te %eratures. H*s t5e 5ig5 te %eratures 'a%ours o'e t5roug5 t5e s$ste : t5ere is dro% in te %erature due to 5eat loss to s$ste and surroundings. 15en t5ese 'a%ours cools to t5eir de0 %oint: li3uid dro%lets are for ed: ..O. "ot is designed for re o'al of li3uid and also ade3uate s%ace is %ro'ided for re o'ed li3uid. * 0ater seal dru is %ro'ided at or near t5e !ase of t5e flare stac- as a safeguard against ingress of air in t5e flare s$ste . *ir u%on entering t5e flare s$ste ixes 0it5 t5e co !usti!le gases 0it5in t5e s$ste and ixture a$ !e 0it5 in t5e ex%losi'e range resulting in so e serious ex%losion. T5ere are nu erous sources of entr$ of air to t5e flare s$ste . 15en t5e flare s$ste is fairl$ static and flo0 is eit5er 'er$ s all or non existent and flares stac- is filled 0it5 eit5er lo0 olecular 0eig5t gases or gases at 5ig5 te %erature t5e draft 0ould exist at t5e !ase of t5e stac-. 1it5 t5is state of less t5an at os%5eric %ressure 0it5 t5e flare s$ste it is o!'ious t5at air 0ill lea- into t5e s$ste if t5ere is o%%ortunit$ for it to do so. T5e 0ater seal %re'ents da age to t5e s$ste u%strea of t5e 0ater seal dru . For %rotection of flare s$ste do0nstrea of t5e 0ater seal: i.e. flare stac- itself: entr$ of air fro t5e flare outlet is %re'ented !$< > Continuous %urge 0it5 ox$gen free gas and > 8se of gas seal. To a'oid a static condition of flo0 t5roug5 t5e flare stac-: continuous %urge is aintained. "urge gases a$ !e an$ ox$gen free gas 05ic5 0ould not reac5 its de0 %oint under an$ condition of flare o%eration. Natural gas: inert gases and nitrogen are co onl$ used: !ut stea is not at all satisfactor$ as a %urge gas. Calculated 'olu es of %urge gas s5ould enter t5e flare s$ste to ensure s0ee%ing of t5e s$ste . T5ere is di'ergent o%inion as to t5e 'olu e of %urge gas 05ic5 a$ !e re3uired. T5e use of gas seals installed i ediatel$ !elo0 t5e flare ti% are ca%a!le to restrict entr$ of t5e air into t5e flare s$ste . Carious t$%es of seals are a'aila!le as discussed !elo0< T5e fluidic seal consists of a series of fixed !affles s5a%ed li-e o%en ended cones @Ref. Fig
> (.2= 0it5in a flare ti%. &ac5 succeeding !affle encountered !$ 0aste or %urge gas tra'eling a%erture t5an !elo0 it. T5e resultant effect is si ilar to t5at of 'enturi 0it5 little dro% in gas %ressure. To force t5e air into t5e flare stac- a!out # to ( ti es t5e %ressure of %urge gas is re3uired and 5ence at os%5eric air can not enter t5e stac- t5roug5 t5e fluidic seal. T5e olecular seal is s5o0n in fig. (.3. T5e %urge gas is forced to a-e t0o 140 o !ends as it flo0s t5roug5 t5e seal. T5e air t5us encounters a tra% caused !$ t5e difference in olecular 0eig5t !et0een air and t5e %urge gas. If t5e %urge gas is lig5ter t5an air: t5e tra% for s in t5e u%%er %ortion of t5e in'erted c$linder. If t5e %urge gas is 5ea'ier t5an air: t5e tra% is for ed in t5e lo0er %ortion of t5e outer c$linder. T5e seal is effecti'e e'en 05en nitrogen is used as a %urge gas. T5e olecular seal is 5ea'ier and large as co %ared to t5e fluidic seal !ecause of t5e co %lexit$ of its internal %assage. T5is necessitates a stronger and costlier su%%ort structure to carr$ t5e additional 0eig5t and 0ind resistance. Bolecular seal o%erates at reduced efficienc$ if its !affles are %artiall$ filled 0it5 0ater or %ieces of refractor$ fro t5e flare ti%. * drain is used for 0ater re o'al: !ut t5e refractor$ re o'al could re3uire costl$ aintenance. Fluidic seal due to its si %licit$ of construction: lig5t 0eig5t and lo0er initial cost: is %referred. ,o0e'er for 'er$ large dia eter flare stac-s @#4I and 5ig5er= t5ere is a !elief t5at olecular seal reduces ox$gen ingress in flare stac-. Flare ti% a %ro%rietar$ ite : ade of 5eat resistant allo$ steel is s%eciall$ designed to %ro'ide a sta!le and 3uiet fla e of t5e desired s o-eless ca%acit$.
..0 +.1
DESI#N OF FLARE S,STE+ Flare L$a!: T5ere are 'arious reasons due to 05ic5 5$drocar!ons andAor ot5er aterials 0ill !e relie'ed to t5e flare s$ste . During eac5 suc5 e ergenc$: its 3uantit$ and is not necessaril$ t5e largest nu !er of .gA,r.: !ut it is t5e flo0 05ic5 i %oses t5e greatest 5ead loss 05ile flo0ing t5ru t5e s$ste . "ressure dro% t5roug5 t5e 5eader is %ro%ortional to t5e s3uare of t5e ass flo0 rate and in'ersel$ %ro%ortional to t5e densit$. It is also %ossi!le t5at during 'arious e ergenc$ conditions different safet$ 'al'es disc5arge and 5ence t5e !ac- %ressure allo0a!le a$ !e different. Flare load fro one %lant< Flare load fro one %lant is esti ated considering< > Flare load > Te %erature and olecular 0eig5t of gases > "er issi!le !ac- %ressure T5e load is esti ated for 'arious e ergencies and to arri'e at design flare load: one e ergenc$ is considered at a ti e in one %lant. Flare Load for a Co %lex< To define t5e s$ste load: t5e si ultaneous occurrence of t0o or ore unrelated contingencies need not !e assu ed. For exa %le: it is generall$ not necessar$ to assu e !loc-ed outlets on %rocess s$ste s under fire conditions. ,o0e'er: under so e arrange ents of %rocess e3ui% ent: a fire could %ossi!l$ result in a failure of local 0iring or instru ent air %i%ing: leading to t5e closure of 'al'es t5at !loc- off t5e %rocess s$ste . &ac5 indi'idual contingenc$ s5ould t5erefore !e re'ie0ed for %ossi!le resultant effects.
"articular stud$ is re3uired for cases of failure of a;or utilities: suc5 as %o0er or cooling 0ater. Co %lete failure of electrical %o0er: cooling 0ater: or stea to an entire %lant s5ould !e considered. 15ere utilit$ sources are !elie'ed to !e unrelia!le or are not !ac-ed u% !$ a s%are unit: t5e effect of co %lete failure s5ould !e studied. T5is t$%e of stud$: 0it5 reference to electrical %o0er failure: co onl$ results in a design !ased on t5e failure of one !us: alt5oug5 loss of an entire distri!ution center or of t5e inco ing line is occasionall$ used as a !asis for design. T5e ca%a!ilit$ of stea s$ste to %ic- u% stand!$ tur!ine loads s5ould !e re'ie0ed in con;unction 0it5 t5e o'erall installed !oiler ca%acit$ and t5e nor al stand!$ ca%acit$ i ediatel$ a'aila!le. T5e ost co on !asis for anal$?ing 0ater or stea failure is t5e failure of one lateral rat5er t5an t5e entire su%%l$. Instru ent air failure is co onl$ considered to !e a %lant 0ide failure unless auto atic a-eu% fro an uninterru%ted source is %ro'ided. Failure of t5e %o0er su%%l$ to electronic or electrical instru ents a$ also !e considered %lant 0ide unless %ro%er stand!$ %o0er su%%lies are %ro'ided. To define t5e co !ined relie'ing loads under fire ex%osure: t5e %ro!a!le axi u extent of a fire s5ould !e esti ated. T5is a$ !e done on t5e !asis of t5e actual la$out of facilities: considering t5e location of sources of co !usti!les: t5e %ro'ision of drainage: and t5e effects of natural !arriers. Facilities t5at 5andle onl$ gaseous fluids a$ !e assu ed to generate ore locali?ed fires t5an t5ose t5at 5andle li3uid co !usti!les. In t5e a!sence of an$ ot5er go'erning factors: consideration of a fire incident is fre3uentl$ li ited to 230 s3uare eters to #+0 s3uare eters. Flare L$a! Re!u ti$n in a C$mple': Flare load in a %lant can !e considera!l$ reduced !ased on follo0ing conce%ts< > Cooling 0ater segregation in %lant so t5at si altaneous cooling 0ater failure in different units can !e a'oided. > )egregation of "o0er Distri!ution t5roug5 'arious )u!>)tations.@ Refer design !asis of ,"CL B !ai.= > Flare load reduction !ased on INT&RLOC.s. @ Refer section 10.0 = @ ex. 15en Cooling 0ater failure occurs in a colu n : stea su%%l$ to col. /otto re!oiler 0ill sto% i ediatel$.= Flare /ea!er: T5e flare 5eader ust !e si?ed so t5at t5e !uilt>u% !ac- %ressure at t5e outlet of relief 'al'es does not exceed t5e axi u %er issi!le 'alue. In so e cases if allo0a!le !ac%ressure is go'erned !$ less load and for a;or flare load 5ig5er !ac- %ressure is allo0a!le t5en t0o %arallel 5eaders one L" and ot5er ," are run if it is econo ical. Flare %i%ing s$ste 1= Indi'idual disc5arge lines fro ")Cs. 2= )u!5eaders in eac5 area connecting to disc5arge lines. 3= Bain flare 5eader leading to .OD. #= Final 5eader connecting t5e 'a%our line to flare stac-. )i?ing Flare ,eaders 1= /ac- %ressure i.e. 10G of set %ressure in %sig for con'entional t$%e and #0G for !alance t$%e 'al'e. 2= To a'oid noise %ro!le : 'elocit$ is li ited. General %ractice @.2 > .(= Bac5 nu !er is selected. )onic 'elocit$ Cc D 223 A .TAB
+.2
Cc in ftAsec . D C%AC' nor all$ T D oR B D Bol. 1t. 15en dia of .OD exceeds 12 ft: s%lit flo0 arrange ent is econo ical For esti ating !ac- %ressure at safet$ 'al'es t5e follo0ing losses are to !e added< )tac- exit loss Gas seal loss @Ta-e % D t5ree 60 o !ends of dia eter e3ui'alent to flare stac-= )tatic 5ead due to stac- 5eig5ts &ntr$ loss to stac&ntr$Aexit loss to 0ater sealdru )eal 5eig5t in 0ater seal dru &ntr$Aexit loss to ..O. Dru Flare 5eader dro% T5e e ergenc$ disc5arge is nor all$ at 5ig5 te %eratures and flare 5eader runs an$ 5undred eters. T5is results in dro% in flare gas te % due to 5eat loss to etal and surroundings. In order to esti ate %ressure loss t5roug5 flare 5eaders: t5e total 5eader lengt5 is di'ided into a nu !er of sections and %ressure dro% s5ould !e calculated for eac5 section se%aratel$ for eac5 section. * c5ec- s5ould !e ade to ensure t5at t5e axi u 'elocit$ in t5e flare 5eader is 0ell !elo0 sonic 'elocit$. Generall$ axi u 'elocit$ is @0.# > 0.(= of sonic 'elocit$ is acce%ta!le. )onic 'elocit$ is esti ated as follo0s< )onic 'elocit$ D 36.3 @g-tAB= J ftAsec 15ere:gD *cceleration due to gra'it$ 32.2 ftAsec2 . D c%Ac' for t5e flare gases T D Te %erature of gases oR B D *'g. Bolecular 0eig5t of flare gases. R D 1.4 . R D #+0K F In so e cases corrosi'e gases li-e ,2) or ex%losi'e gases li-e acet$lene are to !e 5andled 05ic5 s5ould not !e connected to t5e co on 5eader. For suc5 gases a se%arate 5eader is run to t5e flare stac- area. * se%arate 0ater seal dru is %ro'ided 0it5 se%arate stacand is ;oined 0it5 t5e ain stac- !elo0 t5e gas seal. +.3 Flare Sta 0 : Flare stac- is located on do0n0ind %ile and at a re ote %lace fro o%erating ?ones.T5e si?ing of t5e flare stac- re3uires t5e deter ination of stac- dia eter: 5eig5t of stac- and flare lengt5. ,ence one si %le a%%roac5 for flare si?ing is descri!ed< ,eig5t of flare stac- de%ends u%on follo0ing< 1= ,eat released !$ t5e flare gas in .1. 2= C5aracteristics of fla e and fla e lengt5 3= & issi'it$ of fla e #= Reduction intensit$ of flare (= GLC of toxic gases Radiation Intensit$ & issi'it$ 'al'e
> >
.1 .3
D2 D x2 K ,2 x D 20 te 20 ftAs D 'elocit$ of esca%e te D ti e of esca%e to a safe location For a 5ig5 te % s$ste : a se%arate su!5eader a$ !e re3uired u%to t5e %oint: 05ere t5e te % dro%s do0n to t5e allo0a!le li it. * 5eat loss of 10 /T8A5rAft2 a$ !e ta-en as t5u !rule for 5eat loss calculations. For lo0 te % s$ste : flare su!5eaders a$ si ilarl$ co !ined into a single lo0 te % flare 5eaders and %i%ed all t5e 0a$ to flare stac-. ,ere: a 5eat gain calculation is done to find out: 05et5er su!5eader can !e connected to ain 5eader or not. )IB"L& *""RO*C, FOR FL*R& )T*C. )ILING /asic data Baterial flo0ing is ,C 'a%our 1 D Flo0 rate @-gA5r= B1 D *'g olecular 0eig5t T D Te %. Of fluid flo0ing @.= L D Co %ressi!ilit$ factor @Generall$ ? D 1.0 is ta-en= "2 D "ressure at t5e ti% d D Flare dia eter @ = @inside= . D Ratio of s%ecific 5eat , D ,eat of co !ustion @.; %er -g= Cal ulati$n $f Diameter Bac5 No. D 3.23 M 10>( @ 1 A @ "2 M d2 == @L MT =A @ .MB1== 0.( Flare stac- dia is generall$ si?ed on 'elocit$ !asis : alt5oug5 %ressure dro% s5ould !e c5ec-ed. Celocit$ u%to 0.( Bac5 a$ !e %er itted for %ea- load and s5ort ter . 0.2 Bac5 is aintained for ore nor al and %ossi!l$ ore fre3uent conditions for lo0 %ressure flares: 05ic5 de%ends on follo0ing criteria< a=Colu e ratio of axi u concei'a!le flare flo0 to antici%ated a'erage flare flo0. != T5e %ro!a!le ti ing : fre3uenc$: and duration of t5ose flo0s. c= T5e design criteria ado%ted for t5e %ro;ect to sta!ili?e flare !urning. ,o0e'er : for 5ig5 %ressure flares: a 5ig5er Bac5 no. can !e ta-en. Cal ulati$n $f flare len1t& T5e 5eat li!erated: E is calculated as follo0s< E D Flo0 rate x 5eat of co !ustion D @1A3+00=M @ ,= Fro Fig. * flare lengt5 L @ = can !e found out. Calculation of Re3uired Flare )tac-< /asis< )ee Fig. C for di ensional reference. )i?ing of flare stac- is !ased u%on t5e effects of radiation. T5e follo0ing e3uation !$ ,a?en and Lud0ig can !e used to deter ine in distance fro a flare to an o!;ect 05ose ex%osure to t5er al radiation ust !e li ited< D D D @ MFME A @#M M.== D in distance fro t5e id %oint of fla e to o!;ect !eing considered
@in eters= D fraction of 5eat intensit$ trans itted F D fraction of 5eat radiated E D 5eat release @lo0er 5eating 'alue= @.1= . D *llo0a!le radiation @.1A 2= For initial calculation a$ !e assu ed 1.0 *ccording to /r?usto0!sui 7 )u er: use of t5e fraction of 5eat intensit$ trans itted: to correct t5e radiation i %act. D 0.26 @ 100 A r = 1A1+ @30.( A D= 1A1+ D fraction of 5eat intensit$ trans itted t5roug5 t5e at os%5ere. r D relati'e 5u idit$: %ercent D D Distance fro flares to illu inated area @ = @ see fig. C= *!o'e e3uation satisf$ ost of flare gases exce%t ,2 7 ,2) 05ic5 !urn 0it5 little or no lu inous radiation. T5e F factor allo0s for t5e fact t5at not all t5e 5eat released in a fla e can !e transferred !$ radiation. Beasure ent of radiation fro fra es indicate t5at t5e fraction of 5eat radiation @radiant energ$ %er total 5eat of co !ustion= increases to0ards a li it: si ilar to t5e increase in t5e !urning rate 0it5 increasing fla e dia eter. For %ea- flare load F D 0.1 nor al flare load F D 0.3 &sti ate fla e lengt5 fro Fig * indicates fla e radiation center !eing at t5e fla e id %oint. * fla e under influence of 0ind 0ill tilt in t5e direction of 0ind is !lo0ing. T5e lateral 0ind effect can !e esti ated fro Fig./ 05ic5 relates 5ori?ontal and 'ertical dis%lace ent of fla e center to t5e ratio of lateral 0ind 'elocit$ to stac- 'elocit$. Fla e lengt5 'aries 0it5 e ission 'elocit$ and 5eat release. Flo0 in @ 3As= Can !e esti ated !ased on ideal gas la0< 8 D 1ind 'elocit$ 8; Flare ti% 'elocit$ 8; D Flo0 @ 3As= @@ Md 2 =A#= *not5er et5od to calculate t5e ti% exit 'elocit$
sonic 'elocit$ D 61.2 @ -M T; A B ; = 0 . ( N 8 ; D ;et Bac5 no. M sonic 'elocit$ Fro Fig. / 7 OAL can !e esti ated
OAL
15ere L is fla e lengt5 Fro ,a?en and Lud0ig e3uation : D is calculated. No0 corres%onding to desired Bac5 No. stac- 5eig5t: , : is calculated as follo0s< ,9 R9 D2 D D D ,KJ R>J O x
@R9= 2 K @,9= 2
R is esti ated as %er design !asis. T5is corres%onds to . at %articular distance fro flare stac-. Fro +.# a!o'e e3uations , is calculated.
axi u
allo0a!le radiation
Si2in1 a 3n$ 0$ut Drum )i?ing a -noc- out dru is generall$ a trial and error %rocess. T5e first ste% is to deter ine t5e dru si?e re3uired for li3uid entrain ent se%aration. Li3uid %articles 0ill se%arate @a= 05en t5e residence ti e of t5e 'a%or or gas is e3ual to or greater t5an t5e ti e re3uired to tra'el t5e a'aila!le 'ertical 5eig5t at t5e dro%out 'elocit$ of t5e li3uid %articles: and @!= 05en t5e 'ertical gas 'elocit$ is sufficientl$ lo0 to %er it t5e li3uid dro%out to fall. T5is 'ertical 5eig5t is usuall$ ta-en as t5e distance fro t5e li3uid surface. T5e 'ertical 'elocit$ of t5e 'a%or and gas s5ould !e lo0 enoug5 to %re'ent large slugs of li3uid fro entering t5e flare. )ince t5e flare can 5andle s all li3uid dro%lets: t5e allo0a!le 'ertical 'elocit$ in t5e dru a$ !e !ased on t5at necessar$ to se%arate dro%lets fro 300 icro eters to +00 icro eters in dia eter. T5e dro%out 'elocit$ @6= of a %article in a strea is calculated using &3uation 30 as follo0s< 8c 8c g D %1 D D D D D 1.1( Pg D@ l >
'
dro%out 'elocit$: @ eters %er second= acceleration due to gra'it$: @6.4 eters %er second %er second=. "article dia eter: @in eters= densit$ of t5e li3uid at o%erating conditions: @-ilogra s %er cu!ic eter= "' D densit$ of t5e 'a%or at o%erating conditions: @-ilogra s %er cu!ic eter= C D drag coefficient @see Figure 20=. T5e !asic e3uation is 0idel$ acce%ted for all for s of entrain ent se%aration. T5e second ste% in si?ing a -noc- out dru is to consider t5e effect an$ li3uid contained int 5e dru a$ 5a'e on reducing t5e 'olu e a'aila!le for 'a%orAli3uid disengage ent. T5is li3uid a$ result fro @a= condensate t5at se%arates during a 'a%or release: or @!= li3uid strea s t5at acco %an$ a 'a%or release. T5e 'olu e occu%ied !$ t5e li3uid s5ould !e !ased on a release t5at lasts 20 to 30 inutes. *n$ accu ulation of li3uid retained fro a %rior release @fro %ressure relief 'al'es or ot5er sources= s5ould !e added to t5e li3uid indicated in ite s @= and @!= to deter ine t5e a'aila!le 'a%or disengaging s%ace. ,o0e'er: it 0ould usuall$ not !e necessar$ to consider t5e follo0ing 'olu es relati'e to 'a%or disengaging in t5e follo0ing situation t5at in 05ic5 t5e -noc- out dru is used to contain large li3uid du %s fro %ressure relief 'al'es fro ot5er sources 05ere t5ere is no significant flas5ing and t5e li3uid can !e re o'ed %ro %tl$. &cono ics of 'essel design in selecting a dru si?e: 05ic5 a$ influence t5e c5oice !et0een a 5ori?ontal and a 'ertical dru . 15en large li3uid storage is desired and t5e 'a%or flo0 is 5ig5: a 5ori?ontal dru is often ore econo ical. *lt5oug5 5ori?ontal and 'ertical -noc-out dru s are a'aila!le in an$ designs: t5e differences are ainl$ in 5o0 t5e %at5 of 'a%or is directed. T5e 'arious designs include t5e flo0ing<
a.
* 5ori?ontal dru 0it5 t5e 'a%or entering one end of t5e 'essel and exiting at t5e to% of t5e o%%osite end @no internal !affling=. !. * 'ertical dru 0it5 t5e 'a%or inlet no??le on a dia eter of t5e 'essel and t5e outlet no??le at t5e to% of t5e 'essel9s 'ertical axis. T5e inlet strea s5ould !e !affled to direct t5e flo0 do0n0ard. c. * 'ertical 'essel 0it5 a tangential no??le. d. * 5ori?ontal dru 0it5 t5e 'a%or entering at eac5 end on t5e 5ori?ontal axis and a center outlet. e. * 5ori?ontal dru 0it5 'a%or entering in t5e center and exiting at eac5 end on t5e 5ori?ontal axis. f. * co !ination of a 'ertical dru in t5e !ase of t5e flare stac- and a 5ori?ontal dru u%strea to re o'e t5e !ul- of t5e li3uid entrained in t5e 'a%or. T5is co !ination %er its t5e use of large 'alues for t5e nu erical constant in t5e 'elocit$ e3uation. T5e follo0ing sa %le calculations 5a'e !een li ited to t5e si %lest of t5e design. Ite s a and !. T5e calculations for ite s d and e 0ould !e si ilar: 0it5 one>5alf t5e flo0 rate deter ining one 5alf t5e 'essel lengt5. T5e nor al calculations 0ould !e used for ite c and d 0ill not !e du%licated 5ere. T5e follo0ing for ula can !e used for si?ing 5ori?ontal dru s for se%aration of #00 %article< 1 D 3+0M D2 R @eL > eG=M BM"A T SJ 05ere: 1 D l!sA5r of 'a%our eL D li3uid densit$ l!sAcu ft eG D gas densit$ l!sAcuft B D ol. 1t. Of t5e 'a%or T D te %erature of t5e 'a% in oR. " D %sia: D D dru dia eter in fit )i ilar ex%ressions are a'aila!le for 'ertical -noc- out dru s. * %ractical for ula for t5e 'a%or 'elocit$ is< C D 0.# M @eL > eG= AeG ftAsec. +.( 4LO5DO5N *U+* * !lo0do0n %u % is %ro'ided to transfer li3uid fro ..O. Dru . T5e %u % ca%acit$ is %ut e3ual to t5e axi u %ossi!le condensation rate during %ea- flaring. ,o0e'er: t5e ini u %u % ca%acit$ s5ould !e ( 3A5r. T5e %u % s5ould 5a'e %ro'ision to start and sto% auto aticall$ t5roug5 le'el s0itc5es on ..O.D. T5e !lo0do0n %u %s selected s5ould 5a'e ini u N"), re3uire ent and .OD s5ould !e ele'ated suita!l$ to eet t5e N"), re3uire ent. T5e condensate in t5e .OD s5ould !e considered as su!cooled for t5e N"), @a'aila!le= calculations. T$%ical N"), @a'aila!le= for t5e condensate %u % is 2.0 and t5e .OD s5all !e ele'ated at .( @ in= fro t5e grade. ,ig5er ele'ation of t5e .OD results in increased ele'ation of t5e flare 5eader: t5us increasing t5e s$ste costs. 5ater Seal Drum )eal is %ro'ided at t5e !ase of flat stac- to %re'ent an$ flas5 !ac-. In a!sence of seal a continuous entr$ of gas a$ !e !led to flare. )eals are t0o t$%es
+.+
> Li3uid seal > Gas )eal Li3uid seals are furt5er classified a )eal Dru 7 )eal %i%e. In seal dru li3uid seal is located. "i%e is an integral %art of t5e stac-. T5e %ur%ose of seal dru is to aintain a seal of se'eral inc5es on inlet flare 5eader: not exceeding six inc5es ot5er0ise: it a$ cause !ac- %ressure on .O"D. 1ater is nor all$ used on sealing li3uid in cold cli ate so e extra %ro'ision is needed eit5er 0ater is 5eated or 0ater is re%laced !$ alco5ol: -ero etc. For designing 0ater seal dru : )5ell et5od is used. Refer attac5ed fig. +.2 *UR#E #AS RE6UIRE+ENT It is essential to %re'ent air ingress into t5e flare s$ste . T5e ost acce%ta!le safet$ easure is to ensure a %ositi'e flo0 t5ru flare s$ste . It is nor al %ractice in industries to su%%l$ gases to t5e flare s$ste constantl$ to a'oid a static condition of flo0. T5ese gases are referred to as %urge gases. T5e 3uoted 'alues of t5e ini u re3uired %urge gas 'elocit$ 'ar$ fro as lo0 as 0.0+ to 3 ftAsec. ,o0e'er: t5ere is general agree ent t5at a co erciall$ a'aila!le gas seal installed i ediatel$ !elo0 t5e flare at t5e to% of t5e stac- esta!lis5es %erfect safet$ to t5e flare s$ste 05en t5e %urge 'olu e ad itted is ca%a!le of aintaining stac- 'elocit$ fro 0.0( to 0.10 ftAsec. To !e safer side it is reco ended to use %urge gas to 5a'e 0.1 ftAsec: stac- 'elocit$. *ur1e 1a% requirement in t&e flare &ea!er: 1it5 seal 0.1 ftAsec 1it5out seal 1 ftAsec +.4 S+O3ELESS O*ERATION OF FLARE /ecause of increasing stringent air %ollution la0s: s o-eless flare o%eration is %rocess %lants 5as !een %referred in t5e interest of good %u!lic relations. In t5e flaring of gaseous 5$drocar!ons: t5e tendenc$ for s o-e %roduction as t5e gases !urn is go'erned !$ t5e 0eig5t ratio of 5$drogen to car!on in t5e gases !ut is not directl$ %ro%ortional to t5e ,AC ratio !$ 0eig5t. T5ere 5a'e !een nu erous a%%roac5es to t5e %ro!le of s o-eless !urning of 0aste %rocess gases. )ince t5e %rinci%le eans for su%%ression of s o-e in'ol'es t5e 'arious c5e istries associated 0it5 reaction of 0ater 'a%ours 0it5 t5e co %onents of t5e flared gas strea : t5ere 5a'e !een 'arious sc5e es for deli'er$ of t5e 0ater to t5e !urning ?one as eit5er stea or as 0ater s%ra$ in one for or anot5er. T5ere 5as !een certain li ited success 0it5 0ater in;ection for s o-e su%%ression due to t5e great nuisance of a s%ra$ of un'a%ori?ed 0ater to grade in nor al o%eration of t5e flare. Due to t5is reason stea is al ost uni'ersall$ used for su%%ression of s o-e in flare o%eration. 8se of stea increases t5e !urning rate !$ t5e creation of tur!ulence in t5e reaction gases and t5e ins%iration of air: t5ere!$ reducing t5e for ation of soot. )oot for ation is also reduced !$ t5e 0ater gas reaction C D ,2O D CO K ,2: %ro oted !$ t5e addition of stea . T5e a ount of stea for s o-eless o%eration can !e co %uted fro < 1 stea D 1 ,C @0.+4 > 10.4AB= 15ere 1 stea D )tea rate l!sA5r 1 ,C D ,$droca!ron rate: l!A5r
B D Bolecular 0eig5t of 5$drocar!on It is nor all$ not %ossi!le to attain s o-eless o%eration of flare at %ea- flaring loads: as t5e stea rates 0ould !e %ro5i!iti'el$ 5ig5. T5e nor al %ractice is to design t5e flare s$ste for s o-eless o%eration for an$ of t5e follo0ing reasons. 1. Nor al flaring loads 2. )tart>u% or s5utdo0n of a %rocess %lant: or 3. For an$ ot5er situation 05ic5 in'ol'es %rolonged flaring. Flares for su%%ression of s o-e are co erciall$ a'aila!le of %ro%rietar$ designs.: T5e anufacturer s5ould !e consulted on t5e ini u necessar$ stea rate. T5e ste for s o-eless consu %tion can !e %ut t5roug5 a anual re ote control 'al'e or t5roug5 an auto atic s o-e detector: located at t5e ground le'el: sending signal to t5e stea control 'al'e. Total flare load is su of go'erning loads fro different units !ased on go'erning factor of res%ecti'e units. 2.0 IN/OUSE SOFT5ARE A)AILA4LE FOR FLARE S,STE+ SI7IN# Follo0ing %rogra es are a'aila!le in LOT8)>123 in dir D<TFLRUB*N for 'ariouscalculations. Na e of file Function sta-Ucal.0-# -odUcal.0-# %Udro%.0-# TUdro%.0-# *%art fro "ROD&). )tac- dia calculation : %ress dro% fro flare ti% to .OD ,ori?ontal .OD si?ing "ressure %rofile calculation Te %. "rofile calculation a$ refer to 'arious %ac-ages a'aila!le on ain
In a large c5e ical %lant or Refineries 5a'ing a 'er$ relia!le %o0er su%%l$. Cooling 0ater failure ostl$ go'erns t5e axi u flare load. In t5e follo0ing exa %le: s$ste 5as !een segregated into t0o %arts. Cooling To0er one su%%l$ C1 to 8>01: 8>02: 8>03: 8>0# and CT>2 su%%lies to 8>0(: 8>0+: 8>04. Cooling 0ater failure is t5e go'erning case in eac5 unit. Baxi u fire load co es fro CT>1 failure. )o far design of flare 5eader 7 stac- axi u load corres%onding to CT>1 is t5e go'erning load. *ccording to go'erning load: flare stac- dia and .OD 5a'e !een si?ed. Indi'idual unit location is s%ecified !ased on %lot %lan. *ccordingl$: unit !ac- %ressure is calculated o'ing fro .OD to indi'idual units. Carious 5eaders 7 su!>5eaders si?es are o%ti i?ed eeting t5e unit !ac- %ressure re3uire ent. 4.0 Sample Cal ulati$n seets are attac5ed 5ere0it5.
6.0
,"CL @B= "o0er su!station B))>( failure is t5e go'erning case for Flare s$ste design. Go'erning load D ##6 T", Baxi u allo0a!le !ac%ressure at unit /AL D 0.+1 -gAc 2g Bax. *llo0a!le !ac- %ress at ")C outlet D 0.21 -gAc 2g Flare 5eader #4I @after 1)D (#I= )tac- 5eig5t D 100 ) o-eless flaring ca%acit$ D 2 TonsA5r 1ater seal dru is an integral %art of flare stac-.
IOCL @"ani%at= Cooling 0ater failure is t5e go'erning case for flare 5eader si?ing. Onl$ single largest local %o0er failure is considered for c5ec-ing t5e flare 5eader si?ing. Go'erning load D #(2 T", Bax allo0a!le !ac- %ressure D 1.( -gAc 2g at unit /AL Bax allo0a!le !ac- %ressure D 1.2 -gAc 2g at ")C outlet Flare ,eader (#I )tac- 5eig5t D 100 Lengt5 of Flare 5eaderD 3 ) o-eless flaring ca%acit$< 1(G of axi u load 1ater seal dru is an integral %art of flare stac-.
BR"L>II For ain flare 5eader si?ing and flare stac- design: general electric %o0er failure is considered. Go'erning load D 2(0 T",. Bax allo0a!le !ac- %ress. *t unit /AL D 1.0 -gAc 2g Bax allo0a!le !ac- %ress at ")C outletD1.2 -gAc 2g Flare 5eader (#I )tac- 5eig5t D 100 ) o-eless flaring ca%acit$ D 110 T", ,ori?ontal 0ater seal dru se%arate fro flare stac-.
"lant
Te % o C 133
)tacDiaA,t @ A = 1.2A100
B1
Re ar-s
(0>100 1.32A 4A#.( 100 112 1.32A100 2.2A#.4 40 (0 1 32A100 N*A#.( 0.+1A100 2.2A2.#
10. 10.1
FLARE S,STE+ INTRODUCTION Flare s$ste s5all !e installed to !urn t5e fla a!le: toxic or corrosi'e 'a%ors safel$ 05ic5 is disc5arged during startu%: s5utdo0n: nor al o%eration and e ergenc$ cases. Dedicated flare s$ste for "5ase>2 is %ro'ided.
10.2
10.2.1 )$ste
T0o flare collecting s$ste s are %ro'ided. One is ain @L"= flare 5eader and t5e ot5er is sour flare 5eader. ,$drocar!on and sour flares released fro t5e follo0ing units are disc5arged to t5e ain @L"= flare 5eader. > > > > > > > @Note=< CD8ACD8AN)8 Cis!rea-er unit Berox units ,$drocrac-er unit ,$drogen unit GO>,D) unit CCRAN,T unit @future= ," flare fro ,$drocrac-er unit 05ic5 is disc5arged inde%endentl$ in %5ase I s5all !e integrated 0it5 ain @L"= flare 5eader in %5ase 2. t5e follo0ing units are disc5arged to t5e sour flare 5eader.
* ine treating unit )our 0ater stri%%ing unit )ulfur reco'er$ unit ain flare -noc-out dru 05ic5 is located near
T5ese t0o flare 5eaders are connected at t5e flare stac-. 10.2.2 Design Flare Load
For flare s$ste design: follo0ing cases are studied. Flare load for t5e 'arious cases is su ari?ed in t5e attac5ed s5eet. > > > > > Cooling 0ater failure General electric failure Fire case /loc-ed outlet Reflux failure
>
For t5e ain flare 5eader si?ing and flare stac- design: general electric %o0er failure case is considered. Indi'idual flare outlet fro eac5 units are si?ed as listed !elo0. CD8ACD8 C/8 CCR K N,DT GOK,D) B&ROV ,C8 *T8 )1) )R8 ,$drogen > > > > > > > > > > *s %er &IL "ac-age. @General electric failure= *s %er "5ase>I @Reflux failure= General electric failure Fire case K & ergenc$ de%ressure Fire case *s %er "5ase>I @Fire case K & ergenc$ de%ressure= Fire case General electric failure *s %er %5ase>1 @Fire case= *s %er %5ase>1 @")* tri% case=
"ressure %rofile and 'elocit$ for t5e flare 5eader design s5all !e as follo0s. > > > > 0.3 -gAc 2g at ain flare .O Dru located near flare stac-. 1.0 -gAc 2g at eac5 unit /L 1.2 -gAc 2g at eac5 )C !ac- %ressure Celocit$ is less t5an Bac5 No. 0.#.
"ressure %rofile for t5e ain @L"= flare 5eader at general %o0er failure case and for t5e sour flare 5eader at *T8 fire case is s5o0n in t5e attac5ed s5eets. 10.2.3 .noc- Out Dru Fi'e @(= Nos. Of Flare .noc- out dru s are %ro'ided in %5ase>2 as %er follo0ing grou%ings< 1= .noc- out dru > > > in C/8 area is recei'ing flare gas fro
.noc-out dru in CD8ACD8AN)8 area .noc-out dru for ,C8 .noc-out dru in )1) area is recei'ing flare gas for > > > *T8 )1) )R8 is recei'ing flare gas fro all of a!o'e -noc- out dru s. is
(=
Flare -noc- out dru s are si?ed as %er *"I R" (21. Design load for eac5 .O dru su ari?ed as !elo0< .noc- out dru In C/8 *rea For ,C8 In )1) area Bain -noc- out dru 10.2.# Flare )tacFlo0 rate -gA5r 146+4# 104140 4++00 2#122# Bol. 1t. 32.(# 4.4 (0 #3 Te %. oC 12(.# 303 123 133 Case General %o0er failure Fire K & ergenc$ de%ressure Fire case in *T8 General %o0er failure
Flare stac- s5all !e designed !ased on t5e general %o0er failure case of 2(0 tA5. Dia eter and 5eig5t of t5e flare stac- s5all !e deter ined considering t5e ground le'el 5eat intensit$ in accordance 0it5 *"I R" (21. Baxi u 5eat intensit$ at ground le'el s5all !e less t5an #:000 -calA 2.5r excluding solar radiation. Flare ti% s5all !e designed considering s o-eless ca%acit$ of 110 tA5 05ic5 is corres%onding to 1(G of axi u flare load. 10.3 +ATERIAL OF CONSTRUCTION > )our flare 5eader > Bain @L"= flare 5eader > *ll t5e -noc-out dru > 1ater seal dru e%ox$ ce ent coating > Flare stac< < < < < )tress relie'ed -illed car!on steel 0it5 + )tress relie'ed car!on steel 0it5 3 C* )tress relie'ed -illed car!on steel 0it5 + )tress relie'ed -illed car!on steel 0it5 + )tress relie'ed car!on steel 0it5 3 C* C* C* C* and
FLARE LOAD SU++AR, 9*/ASE8II REFINER,: 8nit Na e Cooli ng 1ate r "u %s Rate B 1 .gA5 202: (0. 240 + Gener al &lectr ic Failur e Rate B 1 -gA5 243:+ +6. 30 3
Crude Distillation 8nit @W#1000= Cacuu Dist. 8nit @W#11000= Cis!rea-er 8nit @W#1200= Cis!rea-er 8nit @W(1200= )ulfur Reco'er$ 8nit @W#1#00= )ulfur Reco'er$ 8nit @W(1#00= L"G Berox @W#1(00= .eroseneAFet Berox @W#1+00= ,$drocrac-er @Rx )ect= @W#2000=
Te % o C 41
Te % o C 122
B1 Te % o C 22 126
B 1 +4
Te % o C 202
Rate -gA5
B 1
Te % o C
Rate -gA5
B 1
Te % o C
&;ect stea -gA5r 100 oC: B CD8 7 N CD8 *ccording 002= 16:21 # 16:21 # +1 +1 24# 24#
10:+ 03 10:+ 03
22: # 22. #
112 112
1:+40 1:+40
14 14
1(+ 1(+
212 212 1#(A 133 1#(A 133 66 3## #22 4:31 + 2.1 1#+
Ca%acit$ s
12:#0 4
136
+0:01 0
,$drocrac-er @Frac.)ec= @W#2100= "ol$ *ro *dsor = W#2200= * ine Treating @W#2200= )our 1ater )tri%%er @W#2+00= ,$drogen %lant @W#2300= Gas Oil ,D) @W#((00= Na%5t5a Berox @W#(200=
12:# 0#
#6
21
222:2 21
23
10+
42:4( # #2:23 #
1#2 140 (0 14
3304 2
(2. 2
114
#(#1( 22. 6
102
16:# (0
14
4+:+0 0 2:102
#2:# 00
32
112
/ased on "
/ased on "
4:00 0 1#:6 #6
10 0 30. +
222
1(2
#2.4 126 &xternal Fire Total 632:2#6 -gA5 General "o0er failure #3.0 2#1.22# -gA5r B1 Te %.
133 oC C1 failure total 31#:212 Local %o0er failure of ,OC 244.3#6 -gA5
FL*R& )O)T&B 6.1 INTROD8CTION T5e flare s$ste 0ill !e %ro'ided for safe dis%osal of co !usti!le: toxic gases 05ic5 are relie'ed fro %rocess %lants and offsites during start>u%: s5utdo0n: nor al o%eration or in case of an e ergenc$ suc5 as< o o o o o o o 6.2 Cooling 0ater failure "o0er failure Co !ined cooling 0ater and %o0er failure Instru ent air failure &xternal fire /loc-ed outletAan o%en inlet *n$ ot5er e ergenc$
8NIT C*"*CITO T5e flare loads fro s$ste < t5e follo0ing %lants 5a'e !een considered for t5e design of flare "rocess 8nit Ca%acit$: BBT"* 1. 2. CD8 +.0 CD8 3.3 3. #. (. +. 2. 4. 6. 10. 11. 12. 13. .erosine Berox )R L"G Berox FCC L"G Berox FCC Gasoline Berox )ul%5ur /locFCC8 0.2 ,2 8NIT 0.034 CR8 0.( C/8 0.# //8 0.( O,C8 1.2
6.3
FL*R& LO*D)
6.3.1 CD8ACD8 For t5is 8nit t5e flare loads are gi'en !elo0< Cooling 1ater )$ste Local "o0er Failure Fire Case @*t colK )ta! K BTO stri% Ref Dru = Load -gA5r 112##2
Te % o C 41.06
Te % o C 1#1
T5ese loads 5a'e !een ta-en fro t5e %ac-age. 6.3.2 O,C8 T5e flare loads for t5is unit is gi'en !elo0<
Te o C 2#0
& erg enc$ De%res suri?ati on Load Bol -gA5r 0t 12++1 3.+2 2
Te o C ((
6.3.3 FCC8 Cooling 1ater )$ste Load Bol -gA5r 0t. #0000 #0.0 Local "o0er Failure Load -gA5r 60000 Fire Case Bol 0t. Te % o C (0 210 Load -gA5r #0000 Bol 0t. 32+ Te % o C #22
Te % o C 110
t5e %reli inar$ esti ates a'aila!le for t5e licensor. 5$drogen unit flare load su ar$ for G," and t5e
6.3.# ,2 8NIT T5e loads gi'en !elo0 are ta-en fro flare loads are %rorated. Cooling 1ater )$ste Load Bol -gA5r 0t. +1231 13.+( Local "o0er Failure Load -gA5r
Te % o C (0
T5e flare load for different contingencies s5all !e u%dated 05en data fro unit licensor BAs.: ,T*) is a'aila!le. 6.3.( CR8 T5e data for t5e follo0ing contingencies is ta-en fro t5e %ac-age< Cooling 1ater )$ste Load Bol -gA5r 0t. 20000 23.( 6.3.+ C/8 T5e follo0ing are t5e flare loads gi'en for Cis!rea-er 8nit and are ta-en fro %ac-age. Cooling 1ater Local "o0er Fire Case Local "o0er Failure Load -gA5r 10(000 Fire Case Bol 0t. Te % o C 43.4 133 Load -gA5r (1210 Bol 0t. #0
t5e 5$drogen
Te % o C 2#
Te % o C 10(
t5e %rocess
Failure Load -gA5r 12100 Bol 0t. Te % o C (4 166 Load -gA5r 236+0 Bol 0t. +# Te % o C 12(.4
T5e offgases fro t5is unit s5all !e incinerated in t5e incinerator %ro'ided near t5e unit. T5is unit s5all not !e connected to t5e flare s$ste . 6.3.4 B&ROV T5ese unit are not considered to add>on to t5e flare load in t5e contingencies considered a!o'e. T5e flare load for t5ese units 0ill !e furnis5ed 05en data fro t5e res%ecti'e licensors are a'aila!le. Cooling 1ater )$ste Load Bol -gA5r 0t. ,old Local "o0er Failure Load -gA5r ,old Fire Case Bol 0t. Te % o C Load -gA5r ,old Bol 0t. Te % o C
Te % o C
6.3.6 )8L",8R /LOC. T5e flare loads gi'en !elo0 for t5is !loc- for different contingencies are as %er t5e esti ates a'aila!le for t5is !loc- @consisting of *R8: )1) unit and sul%5ur unit=. T5e flare gases fro t5e !loc- s5all !e routed to t5e sour flare. T5e esti ated flare loads are as follo0s< *R8. Cooling 1ater )$ste Load Bol -gA5r 0t. 2(000 20.0 Local "o0er Failure Load -gA5r ,old Fire Case Bol 0t. Te % o C Load -gA5r ,old Bol 0t. Te % o C
Te % o C 1(0
)1) 8nit
Cooling 1ater )$ste Load Bol -gA5r 0t. 124#.2 2#.23 )ul%5ur 8nit
Te % o C 4+.4(
Fire Case Bol 0t. Te % o C 14.01 1+0 Load -gA5r (+4( Bol 0t. #+.2 Te % o C 1#0
T5e unit s5all 5a'e its o0n incinerator and s5all not !e connected to t5e flare 5eader.
FL*R& LO*D )8BB*RO Cooli ng 1ater Failur e Load -gA5r 20224 0 +6422 Local "o0er Failure Bol 0t +(.0 # ((.6 Te % oC 41.0 6 64.# 110 (0 2# 26.( 10(00 0 12100 43.4 (4 133 166 (1210 236+0 #0 +# 10( 12(. 4 (( 1#0 124#.2 2#.2 3 4+.4 ( L"G )%5ere )R8 do0n Load -gA5r (01+0 0 1(242 0 60000 Bol 0t. 4(.6 124.# 30 Te % o C 16+ 23# 210 Ot5ers Fire Contingenc$
1. 2. 3. #. (. +. 2.
Bol 0t 3.+2
B&ROV& ) 4. OFF)IT& +2#00 ) @Note>1= 6. *R8 2(000 20 1(0 ,OLD ,OLD )1) 132(0 14.01 1+0 (+4( )ul%5ur 8nit 10. TOT*L@2= #(232 >3# >40 3 Figures in t5is !loc- re%resents go'erning flare load fro t5e res%ecti'e units.
(0 #+.2
Note 1<
In case of fire at BT !ullets or ,$drogen !ullets t5e$ s5all also disc5arge to flare.
6.(
FL*R& ",ILO)O",O
6.(.1 T5e cooling 0ater failure !eco es t5e go'erning load for t5e flare 5eader si?ing. 6.(.2 Onl$ single largest Local "o0er failure s5all !e considered for c5ec-ing t5e flare 5eader si?ing. 6.(.3 For different units t5e General electric failure s5all not !e considered as go'erning load as t5e feed to t5e furnaceAfeed %u %s s5all also cease leading to a lo0 flare load. T5e onl$ exce%tions are O,C8 and CR8: releasing 5ig5 load. ,o0e'er t5is load is not go'erning. 6.(.# T5ere s5all !e t0o se%arate flare 5eaders. One carr$ing t5e disc5arges fro e3ui% ents 5andling 5ig5 ,2) contents 5$drocar!ons i.e. )our flare 5eader @carr$ing reliefs fro t5e sul%5ur !loc-= and t5e second one carr$ing rest of t5e 5$drocar!on reliefs i.e. Bain flare. 6.(.( * axi u allo0a!le !ac- %ressure of 1.( -gAc 2g at unit !atter$ li it s5all !e considered for si?ing t5e flar 5eader. ,o0e'er a axi u allo0a!le !ac- %ressure of 1.2 -gAc 2g s5all !e considered at ")C outlet. 6.(.+ For sour flare a axi u allo0a!le !ac- %ressure of 0.( -gAc 2g s5all !e considered at )ul%5ur /loc- !atter$ li it for )R8 do0n case: 5o0e'er for Local %o0er failure case and fire case a !ac- %ressure of 1.( -gAc 2g s5all !e considered at sul%5ur !loc- !atter$ li it. 6.(.2 &IL 5ad suggested routing all ,2) !earing strea s to t5e sour flare !ut IOCL decided to route onl$ t5e )ul%5ur !loc- reliefs to t5e sour flare 5eader. 6.(.4 ) o-eless Ca%acit$< T5e flare s5all !e designed for a s o-eless ca%acit$ e3ui'alent to 1(G of axi u antici%ated fuel gas %roduction @i.e. 1(G of 12.3 T", ol 0t. 4.6(= i.e. > 2 T",. 6.(.6 Flare 5eader s5all !e %urged !$ all units res%ecti'e unit 5eaders 0it5 fuel gas. aintaining a %urge 'elocit$ of 0.1 f%s in
6.(.10 For onitoring flare losses: etering instru ent 0ill !e %ro'ided in eac5 unit !loc- area and also on t5e ain flare 5eader. T$%e of instru ents 0ill !e decided during "7ID stage. 6.+ T,& )O)T&B T5e ain flare 5eader s5all collect t5e e ergenc$ 5$drocar!on reliefs fro all t5e e3ui% ents @exce%t e3ui% ent disc5arging to t5e sour flare 5eader= and s5all route it to ain flare stac-. T5e si?e of t5is flare 5eader 0or-s out to !e (#I. To ta-e care of t5e s5ort ter ,2) ric5 reliefs t5e flare 5eader s5all !e %ost 0eld 5eat treated and 5ardness of 200 /,N aintained. T5e flare stac- dia eter s5all !e (#I and 5eig5t 100 . T5is Flare s$ste s5all 5a'e its o0n dedicated set of fla e front generators and fla e %ilots. T5e lengt5 of t5is flare 5eader is X3 .B. To -ee% t5e 5eig5t of flare 5eader 0it5in reasona!le li its and to reduce t5e cost of associated flare 5eader structure: it is %ro%osed
to 5a'e one -noc- out dru enroute flare stac-. T5e -noc- out dru s5all !e located near //8. It s5all 5a'e second -noc- out dru near flare stac-. * 'ertical 0ater seal dru s5all !e %ro'ided at t5e !otto of t5e stac-. T5e 0ater seal dru is an integral %art of t5e flare stac-. It is not %ossi!le to ta-e 0ater seal dru under aintenance 0it5out s5utdo0n of entire co %lex. T5us to a'oid an$ suc5 situation all t5e internals of 0ater seal dru s5all !e of ))>30#. T5e sour flare 5eader s5all 5a'e a -noc- out dru in t5e )ul%5ur !loc-. T5e gases 0ill t5en !e routed to t5e flare area. T5e gases s5all !e !urnt along0it5 t5e ain flare gases t5roug5 a se%arate 5eader near t5e ti% of t5e ain flare. )our flare 5eader s5all !e %ost 0eld 5eat treated and s5all 5a'e ini u 200 /,N 5ardness. T5is flare 5eader s5all 5a'e a .OD near flare stac-. T5is 0ill not 5a'e an$ 0ater seal dru . * se%arate fluidic seal: se%arate Fla e "ilots s5all !e %ro'ided for t5is flare. T5e fla e front generators s5all !e co on for t0o flares. * sc5e atic for flare gas collection 5eader is gi'en under *nnexure 1.
4.0 4.1
FL*R& )O)T&B INTROD8CTION *t %resent se'eral safet$ 'al'es fro existing units are 'enting to at os%5ere. T5e existing flare s$ste is inca%a!le of 5andling t5is additional load. * ne0 flare s$ste is en'isaged to 5andle safet$ 'al'e reliefs fro t5e existing units %resentl$ connected to flare as 0ell as 'enting to at os%5ere and D,D) !loc-.
4.2
FL*R& LO*D) )afet$ 'al'e disc5arges for cooling 0ater failure case fro 'arious units are gi'en !elo0< 8nits FR& *") FR& C") ,B8 T>1201 FR>*") FR>C") FCC8 "ro%ane 8nit T>401 @Deet5= T>402 @De!ut= B&* RegenerationM T>201 @"G C"LG= D,D) ,$drogen )1)M *R8M )R8M M )our gases !eing routed t5roug5 a se%arate acid 5eader. Notes< 1. For /loc-ed outletAfire case flare load details refer flare stud$ re%ort @dated )e%t 166+= su! itted to ,"CL > Bu !ai. 3.( 31.( +( 220 2.4 3+.( 22.3 222 2.4 136 (.( 11.+ 1(.2 4( #.( (2 2.2 16.2 20 20.3 20 20 20 20.3 20 3( (0 32.2 (0.( 22.( 20 32 12.32 1(0 1(0 100 1(0 1(0 1(0 100 (0 (0 22.2 +4.1 12( 1(0 2( 3( > > > > > > > > Lean oil failure > > > > >
4.3
GOC&RNING FL*R& LO*D "o0er is distri!uted in different units in t5e fuel refiner$ 7 FR cooling to0ers t5roug5 'arious su!stations 'i?. B))>1: B))># 7 B))>(. *fter B) axi isation: total flare loads en'isaged for B))>1: B))># 7 B))>( failure are as follo0s< B))>1 F*IL8R& B))>1 su%%lies %o0er to FR *"): FR C"): T>201 @"G CL"G= units 7 3 %u %s of FR cooling to0er. T5erefore: B))>1 failure 0ill result t5e %o0er failure in t5e a!o'e said units 7 t5ree FR cooling 0ater %u % tri%. )ince t5ere is onl$ one co on C1& su%%l$ 5eader fro FR cooling to0er: due to tri% of t5ree C1 %u %s t5e %ressure in t5e 5eader 0ill !e so lo0 t5at t5e ot5er %u %s 0ill also tri% leading to FR cooling 0ater s$ste failure. 8NIT) @oC= FR& CT Not affected FR& *") > > affected FR& C") > > affected ,B8 > > affected T>201 @/, CL"G= > > affected FR CT > Not LO*D: @T",= B1 Te %. Re ar-s
> Not
> Not
> Not
FR C1) 0ill tri% FR *") > > > /ot5 C1 7 unit %o0er failure FR C") > > > >do> FCC8 136 #3.2 failure "RO"*N& 8NIT (.( 3( (0 11.+ (0 (0 T>401 @D&&T,= 1(.2 32.2 oil failure T>402 @D&/8T= 4( (0.( failure B&* R&G&N. #.( 22.( 12( T>201 @"G CL"G= (1.( 20 63 C1
22.2 Lean
+4.1 C1
302.4
+SS8< FAILURE B))># su%%lies %o0er to FR *"): FR C"): ,B8: T>1201 @/, CL"G=. "ro%ane units: one %u % in FR cooling to0er. T5erefore: B))>1 failure 0ill result t5e %o0er failure in t5e a!o'e said units 7 all %u %s in FR cooling to0er. T5erefore B))># failure 0ill result in %o0er failure in t5e a!o'e said units 7 FR& cooling 0ater s$ste failure. )ince onl$ one %u % in FR cooling 0ater s$ste 0ill tri% it 0ill not affect t5e FR cooling 0ater s$ste . 8NIT) @oC= FR& CT FR& C1) 0ill tri% FR& *") > > > /ot5 C1 7 8nit %o0er failure FR& C") > > > >do> ,B8 100 20 100 >do> T>201 @/, CL"G= 22.3 20 1(0 >do> FR CT Not affected FR *") > > affected FR C") LO*D: @T",= B1 Te %. Re ar-s
> Not
> > affected FCC8 > > affected "RO"*N& 8NIT > > > T>401 @D&&T,= > > affected T>402 @D&/8T= > > affected B&* R&G&N. > > affected T>201 @"G CL"G= > > affected Total @for ,C flare= 122.3
> Not
> Not
> Not
> Not
> Not
> Not
B))>( F*IL8R& B))>( su%%lies %o0er to FCC8: T>401 @Deet5aniser=: T>402 @De!utaniser=: B&* Regenerator 7 T0o %u %s of FR cooling 0ater s$ste . T5erefore: B))>( failure 0ill result t5e %o0er failure in t5e a!o'e said units 7 t0o FR cooling 0ater %u % tri%. )ince t5ere is onl$ one co on C1 su%%l$ 5eader fro FR cooling to0er: due to tri% of t0o C1 %u %s t5e %ressure in t5e 5eader 0ill !e so lo0 t5at t5e ot5er %u %s 0ill also tri% leading to FR cooling 0ater s$ste failure. 8NIT) @oC= FR& CT Not affected FR& *") > > affected FR& C") > > affected ,B8 > > affected T>201 @/, CL"G= > > affected FR CT FR C1) 0ill tri% FR *") 222 > Not > Not LO*D: @T",= B1 Te %. Re ar-s
> Not
> Not
1(0 C1
1(0 C1
> /ot5 C1 7 8nit %o0er failure "RO"*N& 8NIT (.( 3( (0 11.+ (0 (0 T>401 @D&&T,= 1(.2 32.2 22.2 Lean oil failure T>402 @D&/8T= 4( (0.( +4.1 /ot5 C1 7 8nit %o0er failureM B&* R&G&N. #.( 22.( 12.( T>201 @"G CL"G= (2 20 failure Total @for ,C flare= ##6.1
> >
1(0 C1
M)ince stea 0ill !e on in t5e re!oiler: o'er %ressurisation 0ill !e t5ere. T5erefore load s5ould !e considered. It is e'ident fro flare design. t5e a!o'e load distri!utions t5at B))>( failure is t5e go'erning case for
4.#
"RO"O)&D )O)T&B
4.#.1 ,$drocar!on 7 sour flare 5eaders 0ill !e segregated. )our flare 5eader 0ill !e stea traced. 4.#.2 )e%arate ,AC flare .ODs 0ill !e %ro'ided for FR& 8nit @44>D>01=N Co !ination 8nit @44>D>02=: L&8 @44>D>03=: D,D) 8nit @/$ licensor= and ,$drogen 8nit @/$ licensor=. 4.#.3 * ain ,AC flare .OD @44>D>02= 0ill !e %ro'ided at flare stac- area. 4.#.# * co on sour flare .OD @44>D>04= 0ill !e %ro'ided at t5e flare stac- area. @44>D>06= for ,AC flare 0ill !e %ro'ided. caused !$ oist
4.#.( * 'ertical 0ater seal dru 4.#.+ No 0ater seal dru ,2)AN,3. 4.#.2 Flare stac- 0ill !e
is %ro'ided for sour flare to a'oid corrosion %ro!le ounted on 0ater seal dru .
4.#.6 * 3+I ,AC flare 5eader is ta-en fro FR& 8nit. T5is is ;oined !$ a 30I ,AC flare 5eader fro L&8. * #2I ,AC flare 5eader is ta-en fro co !ination unit 05ic5 ex%ands to #4I ;ust !efore t5e ;oining of 3+I ,AC flare 5eader fro FR& 8nit to it. * 2#I ,AC flare 5eader fro D,D) /loc- is ;oined to t5e #4I ain 5eader running u%to ain flare .OD. It exits t5e .OD as (#I 5eader !efore it enters t5e 0ater seal dru . * 10I sour flare 5eader 0ill !e %ro'ided fro L&8: 05ic5 0ill !e connected to t5e 10I sour flare 5eader fro D,D) unit. T5e flare 5eader runs u%to sour flare .OD. *fter .OD t5e line runs u%to t5e to% of t5e flare stac- t5roug5 fluidic seal. Refer flare s$ste "7IDs for detail.
Bain Flare )$ste T5e flare load for 'arious units is enclosed as *nnexure>1. T5e flare load for t5e controlling case i.e. cooling 0ater 7 %o0er failure is 222 TA5r. *n atte %t 5as !een ade to reduce t5e load. T5e detailed anal$sis for t5e controlling case 5as !een done 0it5 res%ect to instru entation %ro'ided and rationali?ation of flare loads in e'ent of si ultaneous release fro se'eral colu ns in a unit. T5e details are as under< GC8 *s %er )1&C flare load during electrical %o0erAcooling 0ater failure is as follo0s< Flare load: -gA,r B1 Te %. C 1. ,otA1et flare 5eader 31(360 33.2( 2. Inter ediate Flare ,eader +44(( #2.04 3. Lo0 te %erature flare 5eader 4+(03 24.4+ TOT*L LO*D #202#4 33.+2 /eing a co %lex unit 0it5 in!uilt argins for instru ented tri%s: total flare load 5as !een considered for t5e design of flare s$ste . C)8AC2>C3 8unit< Control sc5e e 5as %ro'ided a nu !er of s5utdo0n 'al'es at different locations to close during %o0erAcooling 0ater failure. T5e details are as follo0s< T0o 3uic- s5utdo0n 'al'es @one redundant= on inlet feed gas to 8""C. *ctuation is due to 5ig5 5ig5Alo0 lo0 %ressure and %o0er failure. T5e gas 0ill !e di'erted to lean gas des%atc5 5eader. )5utdo0n 'al'e u%strea of G)8 a!sor!er to close during %o0er failure due to tri%%ing of lean a ine %u %. During %o0er failure: G)8 s5utdo0n 'al'e s5all close and t5is in turn 0ill close t5e feed gas s5utdo0n 'al'e due to u%strea line %ressuri?ation. )5utdo0n 'al'e in C2AC3 %lant on suction line of lean gas co %ressor. T5e s5utdo0n 'al'e 0ill actuate during %o0er failure due to lo0 lo0 %ressure in t5e suction sideA5ig5 5ig5 %ressure in t5e disc5argeA5ig5 5ig5 te % in t5e disc5arge. T5ere is a re ote %ossi!ilit$ of failure of all s5utdo0n 'al'es to actuate during %o0er failure situation. )till t5e %ossi!ilit$ of %ressuri?ation u%to 1# -gAc 2a 5as !een considered. For lean gas suction line nor al o%erating %ressure is 12 -gAc 2a and ")C is set at 1( -gAc 2a. T5e esti ated flare load due to %ressuri?ation is 2 TA,r. For flare design: conser'ati'e figure of ( TA5r 5as !een considered.
o
o o o
,ig5 !oils colu n 1#6+60 3. Lo0 !oils colu ns +#2#2 #. F& colu n #4(0 TOT*L 324+31 T5e flare load fro 'arious colu ns is for t5e conditions 05en unit %o0er 5as failed !ut stea A/F1 is continuous. During co %lex %o0er failure: stea generationA./F1 s5all also tri%. )u%%l$ of stea to unit re!oilers s5all !e li ited to t5e s$ste 5old u%Asurge a'aila!le. 1it5 t5e li ited stea 5oldu% t5e si ultaneous release of 5$drocar!ons fro 'arious colu ns is not en'isaged. For t5e flare design t5e single largest ")C release fro CB Colu n and 10G of !alance ")C release 5as !een considered as follo0s< Flare load: -gA,r B1 Te %. o C 1. CB Colu n 1(64#6 (1.6 46.+ ,/ Colu n 1#6+6 104.# 2#1.0 3. L/ Colu n +#2# (4.1 1#+.0 #. F& Colu n #4( (4.1 134.1 141222 *'g. (#.#+
2.
2.
Total
#U=ARAT REFINER, FLARE S,STE+ T5e existing flare s$ste of Gu;arat Refiner$ is integrating GR: GR&" and GR)"F in a single flare s$ste due to se%arate dedicated cooling 0ater s$ste s. G," is 5a'ing its dedicated flare s$ste . GR flare is connected to G," for flexi!ilit$ of using G," flare s$ste 05en GR flare is under aintenance. Recentl$ IOCL 5as gi'en a stud$ to connect at os%5eric 'ents of 'arious crude units: incor%orating ne0 units 'i? *8>C: D,D): ,2 8nit: )ul%5ur !loc- : BT/& unit and /utene> 1. )e'eral alternati'es 0ere studied. T5e least cost o%tion 5as e erged !$ considering t5e follo0ing !asis< > )tagger GR cooling 0ater s$ste 0it5 dedicated su!stations so t5at cooling 0ater failure of *8>C: D,D): ,2 8nit and sul%5ur !loc- can !e considered se%arate fro existing *8>I: *8>II: CR8: "DF and Food Grade ,exane unit. > 8se G," flare s$ste for GR at os%5eric 'ents etc !$ follo0ing t5e %5iloso%5$ of single cooling 0ater failure at a ti e. > "ro'ision of G," to GR side flexi!ilit$ to ta-e G," flare s$ste under aintenance. > 8se of s5ell %5iloso%5$ !$ considering tri%%ing of source of 5eat 05ene'er %ressure of cooling 0ater su%%l$ 5eader is 'er$ lo0.
11.3
INTROD8CTION &xcess 5$drocar!ons !eing 0asted to flare s$ste can !e reco'ered !$ using a si %le !ut effecti'e flare gas reco'er$ s$ste . )O)T&B D&)CRI"TION Flare gas reco'er$ s$ste in its ost !asic for is si %l$ a co %ressor to co %ress t5e flared gases for its su!se3uent use at fuel gas in t5e refiner$A%etroc5e ical co %lex. T5e o!;ecti'e of t5e s$ste is to reco'er excess 5$drocar!ons !eing flared and also reductionAcontain ent of radiation: lig5t and glare: s o-e and noise in t5e ad;oining areas. T5e critical as%ect of t5e s$ste is selection of a%%ro%riate ca%acit$ for co %ressor and to set its o%erating s%ecifications so t5at it can safel$ and efficientl$ 5andle as 0ide a range of %ossi!le co %ositions and rates as %ractical. *%art fro t5is: designed s$ste s5ould aintain t5e integrit$ of existing flare linesAflare stac- s$ste . To ac5ie'e t5is: flare s$ste ust !e left as it 5ad !een. No 'al'es auto atic or anual s5ould !e %lace din t5e flare lines to di'ert gas fro flare stac- to co %ressor. )uc5 ec5anical de'ices can fail to o%erate 05en re3uired during e ergenc$. &x%erience s5o0s t5at t5e$ 0ill fail e'entuall$ t5en 0e 0ill 5a'e no o%era!le e ergenc$ s$ste . Reco'er$ s$ste s5ould !e controlled so t5at it does not lo0er flare s$ste %ressure to t5at le'el at 05ic5 air could ingress into t5e s$ste . For t5is: suction %ressure of t5e co %ressor is aintained 0it5 co %ressor ca%acit$ control !ac-ed u% !$ disc5arge rec$cle control. D&)CRI"TION OF R&COC&RO )O)T&B T$%ical reco'er$ s$ste is s5o0n in figure>1. "art of flared gases fro ain flare 5eader do0nstrea of stac- area flare gas -noc- out dru are di'erted to co %ressor t5roug5 a co %ressor suction -noc- out dru to %re'ent an$ li3uid 05ic5 a$ condense in t5e inta-e line or entrained li3uid fro !eing %ulled into t5e co %ressor. T5e -noc- out dru is %ro'ided 0it5 ist eli inator to allo0 onl$ li3uid free gases to co %ressor suction. For noise su%%ression: silencers are installed at !ot5 t5e suction and disc5arge ends of t5e co %ressor. *s t5e gases get 5eated u% in t5e co %ressor: t5ese are cooled in t5e co %ressor after cooler !efore !eing led to t5e fuel gas 5eader. T5e li3uid condensate if an$ is re o'ed in t5e -noc- out 'essel. Design of suction and disc5arge -noc- out dru and also of after cooler are si %le and 0ill not %ose an$ %ro!le due to fluctuation in flared gas co %osition and te %erature. Critical area of t5e flare reco'er$ s$ste is t5e selection of co %ressor and t5e sc5e e to aintain suction %ressure.
12.0
Um$e>% $n ept f$r Flare #a% Re $-ery T5e ne0 conce%t of flare gas reco'er$ eli inates t5e need of continuous flaring of gas.
T5e
total 5$drocar!on gas is reco'ered fro t5e flare s$ste and routed to t5e fuel gas 5eader after co %ression. T5e flare line is closed during reco'er$ of t5e flare gas !$ eans of a onAoff 'al'e installed do0nstrea of t5e flare -noc- out dru . For safet$ reasons a !ursting disc is installed in %arallel 0it5 t5e isolation 'al'e. T5e co %ression facilit$ is designed to 5andle nor al gas lea-age rates: 0it5 a s%are ca%acit$ to anage s aller releases for !lo0do0n 'al'esA")Cs. During large releases t5e 'al'e in t5e flare line 0ill o%en @t$%icall$: t5e o%ening ti e for t5e 'al'e is less t5an 2 secs= and t5e reco'ering e3ui% ent 0ill !e isolated. In case of alfunctioning of t5e 'al'e in flare line: t5e ru%ture disc 0ill o%en and t5e gas 0ill !e flared. T5e flare gas ignition s$ste 0ill ignite t5e flare gas. T5e flare gas ignition s$ste consists of a co %ressed nitrogen dri'en launc5er contained in a stainless steel ca!inet at t5e decle'el. In addition a target %late is located !elo0 t5e flare ti%s to ignite t5e s%eciall$ designed ignition %ellet !eing auto aticall$ launc5ed fro t5e launc5ing unit. )tandard launc5ing range 12( : Launc5ing ediu N2: Launc5ing %ressure 1(0 > 220 !ar: Bu??le 'elocit$ #30 Asec @free= A 320 Asec @guided=: ignition %ellet dia 20 A0t > #3 gra =. 15en 5itting t5e target %late: t5e %ellet 0ill ex%lode and generate a s5o0er of s%ar-s: eac5 igniting along its entire %at5. T5e s$ste launc5es t0o %ellets: t5e first 1( seconds and t5e second 30 seconds after t5e o%ening of t5e 'al'e in t5e flare line. &x%erience s5o0s t5e gas is ignited !$ 6 out of 10 %ellets. )$ste Failure )cenario T5e follo0ing t0o ain safet$ scenarios need to !e considered< 1= /ot5 t5e flare 'al'e and t5e ru%ture disc fail to o%en. 2= T5e flare gas is not ignited. )cenario @1= 5as an extre el$ lo0 %ro!a!ilit$ of occurrence: sufficientl$ lo0 to !e ignored.T0o ru%ture disc in %arallel can !e used to lo0er its %ro!a!ilit$ furt5er. )cenario @2= is not uc5 different fro a continuousl$ !urning flare 05ere a sudden e ergenc$ release or strong 0inds can extinguis5 t5e fla e. ,o0e'er: t5e safet$ of a'iation traffic s5ould !e considered %articularl$ and %rocedures in t5is context: s5ould !e ade. T5is ne0 conce%t of flare gas reco'er$ reduces t5e e ission of Nox 7 CO2 to a large extent. T5is tec5nolog$ 5as !een installed in se'eral %ro;ects in Nor0a$ and %atented !$ BAs. 8BO& Oil 7 Gas.