1991 PUBL 2028 Flame Arresters in Piping Systems

Flame Arresters in Piping Systems API PUBLICATION 2028 SECOND EDITION, DECEMBER 1991 American Petroleum Institute 1220 L Street, Northwest Washington, D.C. 20005

Flame Arresters in Piping Systems Safety and Fire Protection Department API PUBLICATION 2028 SECOND EDITION, DECEMBER 1991 American Petroleum Institute

SPECIAL NOTES
1. API PUBLICATIONS NECESSARILY ADDRESS PROBLEMS OF A GENERAL NATURE. WITH RESPECT TO PARTICULAR CIRCUMSTANCES, LOCAL, STATE, AND FEDERAL LAWS AND REGULATIONS SHOULD BE REVIEWED. 2. API IS NOT UNDERTAKING TO MEET THE DUTIES OF EMPLOYERS, MANUFACTURERS, OR SUPPLIERS TO WARN AND PROPERLY TRAIN AND EQUIP THEIR EMPLOYEES, AND OTHERS EXPOSED, CONCERNING HEALTH AND SAFETY RISKS AND PRECAUTIONS, NOR UNDERTAKING THEIR OBLIGATIONS UNDER LOCAL, STATE, OR FEDERAL LAWS. 3. INFORMATION CONCERNING SAFETY AND HEALTH RISKS AND PROPER PRECAUTIONS WITH RESPECT TO PARTICULAR MATERIALS AND CONDITIONS SHOULD BE OBTAINED FROM THE EMPLOYER, THE

MANUFACTURER OR SUPPLIER OF THAT MATERIAL, OR THE MATERIAL SAFETY DATA SHEET. 4. NOTHING CONTAINED IN ANY API PUBLICATION IS TO BE CONSTRUED AS GRANTING ANY RIGHT, BY IMPLICATION OR OTHERWISE, FOR THE MANUFACTURE, SALE, OR USE OF ANY METHOD, APPARATUS, OR PRODUCT COVERED BY LETTERS PATENT. NEITHER SHOULD ANYTHING CONTAINED IN THE PUBLICATION BE CONSTRUED AS INSURING ANYONE AGAINST LIABILITY FOR INFRINGEMENT OF LETTERS PATENT. 5. GENERALLY, API STANDARDS ARE REVIEWED AND REVISED, REAFFIRMED, OR WITHDRAWN AT LEAST EVERY FIVE YEARS. SOMETIMES A ONETIME EXTENSION OF UP TO TWO YEARS WILL BE ADDED TO THIS REVIEW CYCLE. THIS PUBLICATION WILL NO LONGER BE IN EFFECT FIVE YEARS AFTER ITS PUBLICATION DATE AS AN OPERATIVE API STANDARD OR, WHERE AN EXTENSION HAS BEEN GRANTED, UPON REPUBLICATION. STATUS OF THE PUBLICATION CAN BE ASCERTAINED FROM THE API AUTHORING DEPARTMENT [TELEPHONE (202) 682-8000]. A CATALOG OF API PUBLICATIONS AND MATERIALS IS PUBLISHED ANNUALLY AND UPDATED QUARTERLY BY API, 1220 L STREET, N.W., WASHINGTON, D.C. 20005. Copyright 1994 American Petroleum Institute

FOREWORD
This publ icati on is inte nde d to alert indu stry to the limi tatio ns of flam e

arre sters . Fla me arre sters are usua lly teste d in conf igur atio ns that prod uce low flam e spee ds only . The high ly vari able con ditio ns surr oun ding arre ster appl icati ons may resu lt in high

flam e spee ds that coul d rend er the arre sters inef fecti ve. API publ icati ons may be used by any one desi ring to do so. Eve ry effo rt has bee n mad e by the Insti tute to assu re the

accu racy and relia bilit y of the data cont aine d in the m; how ever , the Insti tute mak es no repr esen tatio n, warr anty , or guar ante e in con nect ion with this publ icati on and here by expr essl y disc

laim s any liabi lity or resp onsi bilit y for loss or dam age resu lting fro m its use or for the viol atio n of any fede ral, state , or mun icip al regu latio n with whi ch this publ icati on may

conf lict. Sug gest ed revi sion s are invit ed and sho uld be sub mitt ed to the dire ctor of the Safe ty and Fire Prot ecti on Dep artm ent, Am eric an Petr oleu m Insti tute, 122 0L Stre et, N.

W., Was hing ton, D.C . 200 05.

SECTION 1 - GENERAL 1.1 Purpose

The avai labil ity of com mer cial flam e arre sters liste d by nati onal ly reco gniz ed testi ng labo rato ries has freq uent ly led to the inst

allat ion of thes e arre sters in pipi ng syst ems ; how ever , the actu al con ditio ns und er whi ch thes e arre sters will oper ate may be far diff eren t fro m the con ditio ns und er

whi ch the arre sters wer e teste d and liste d by the testi ng labo rato ries. List ed arre sters are nor mall y teste d und er nonf lowi ng con ditio ns and with ignit ion at the ope n ends of

pipe s atta che d to the arre sters . The testi ng is con duct ed with limi ted leng ths of the atta che d pipe s. Fla mes prop agat ing thro ugh pipi ng syst ems cont inuo usly acce lerat e and

can reac h velo citie s that are muc h high er than thos e at whi ch the arre sters wer e teste d. As a resu lt, flam e arre sters , whe ther liste d or not, may not be effe ctiv e whe n they

are inco rrect ly appl ied in pipi ng syst ems. List ed arre sters sho uld not be inst alle d in pipi ng syst ems unle ss they hav e bee n teste d und er con ditio ns equi vale nt to thos e exp

ecte d in the spec ific appl icati ons. This publ icati on is inte nde d to alert indu stry to the limi tatio ns of flam e arre sters . The high ly vari able con ditio ns surr oun ding appl icati ons of arre

sters may resu lt in high flam e spee ds that coul d rend er the arre sters inef fecti ve.

1.2 Referenced Publications

The mos t rece nt editi ons of the foll owi ng stan dard s, cod es, and spec ifica tion s are cite

d in this publ icati on. API Publ 2210 Flame Arrester for Vents of Tanks Storing Petroleum Products UL1 UL 525 Flame Arresters for Use on Vents of Storage Tanks for Petroleum Oil and Gasoline Gas and Oil Equ ipm ent

SECTION 2 - FLAME PROPAGATION

Fla mes prop agat ing thro ugh pipi ng syst ems are capa ble of reac hing extr eme ly high

spee ds. Initi ally, the flam es trav el at a burn ing velo city char acte risti c of the mixt ure; this velo city, som etim es tabu late d in han dbo oks, is usua lly a few feet per seco nd. The n the flam es

begi n to acce lerat e. This acce lerat ion proc ess is assi sted by turb ulen ce, whi ch can be indu ced in the unb urne d mixt ure by the flam es the mse lves or can resu lt fro m such fact

ors as flow , pipe wall roug hnes s, or turb ulen ceprod ucin g fitti ngs [1]. Note: The flame velocity at a given point is a function of the length of pipe and size of pipe through which the flames have traveled, the intensity of turbulence, the properties of the particular flammable mixtures and other factors. Fla mes can acce lerat e to a velo city that per mits their trav el upst rea m as well as dow nstr

eam of the orig inal dire ctio n of the flow [2]. The y can read ily reac ha velo city of seve ral hun dred feet per seco nd. If the flam es are prop agat ing in the unb urne d med ium at a velo

city less than the spee d of sou nd, it is kno wn as a defl agra tion, but if the pipe is long eno ugh, flam e prop agat ion und er deto nati on con ditio ns can occ ur. In deto nati ons, flam es can

trav el seve ral thou sand feet per seco nd and are acco mpa nied by pres sure puls es; the mag nitu de of the pres sure puls es may exce ed 20 time s the initi al abso lute pres sure [1]. Alth oug h

the mag nitu de of the pres sure puls es may exce ed 20 time s the initi al abso lute pres sure , ther e are a num ber of way s in whi ch eve n high er pres sure s can be gen erat ed:

a. At closed ends and elbows, the pressure is increased by reflection of the detonation wave. b. At the point where the deflagration transforms into a detonation, even higher pressures can occur during a phase of the detonation known as the overdriven phase [3]. c. When flames are propagating toward a closed system, pressures higher than 20 times the initial absolute pressure can occur under certain circumstances because of a process known as pressure piling. In pressure piling, the deflagration causes precompression of the gas before the transition to detonation. Fla me prop agat ion mak es the inst allat ion of flam e arre sters in pipi ng syst ems fund ame ntall y diff eren t fro m the inst allat

ion of arre sters on tank vent s. Tan k vent s hav e little or no pipe leng th pres ent bet wee n the arre ster and an exte rnal ignit ion sour ce at the ope n end (see Und erwr iters Lab orat

orie s' Gas and Oil Equ ipm ent and API Publ icati on 221 0).

SECTION 3 - LISTED FLAME ARRESTERS 3.1 Problems in Piping Systems

Inst allin g flam e arre sters in pipi ng syst ems with exte nde d or cont inuo us pipe leng ths pres ents a com

plex desi gn prob lem. The crite ria for the desi gn of flam e arre sters for cont inuo us pipi ng syst ems hav e not eve n bee n esta blis hed yet. Syst ems with exte nde d or cont inuo us pipe

leng ths may hav e pipe leng ths that wou ld per mit flam es to attai n eno ugh velo city to pass thro ugh an arre ster. Alte rnati vely , the high pres sure s dev elop ed may dam age the arre stin g

ele men t or rupt ure the hou sing , ena blin ga flam e to pass thro ugh the devi ce. Pres sure s resu lting fro m with in a pipe may exce ed the stre ngth of an atta che d vess el.

3.2 Design

List ed flam e arre sters are usua lly cell ular arre sters and incl ude the foll owi ng type s: perf orat edplat e arre sters , para llelplat e arre sters , crim pedmet alribb on arre sters ,

and sint ered met al arre sters [4]. The se devi ces are barr iers that arre st the flam es by que nchi ng (that is, the heat of the flam es is tran sferr ed to the wall s of an arra y of sma ll

pass age way s in the arre ster) . The criti cal para met ers that gov ern the effe ctiv enes s of cell ular arre sters are the dia met er or widt h and the leng th of the flam e pass ages . Cell ular arre

sters (dis cuss ed in API Publ icati on 221 0) are liste d by nati onal ly reco gniz ed testi ng labo rato ries as inte nde d for inst allat ion on atm osp heri c pres sure tank vent s stori ng petr

oleu m prod ucts and on vent s fro m the tank s of oil tank ers (see Gas and Oil Equ ipm ent) [1]. The listi ngs are usua lly base d on tests mad e with mixt ures of gaso line vap or and air and

may not cov er othe r mixt ures . Arre sters teste d with gaso line vap or are prob ably suit able for use with mos t com mon para ffin or aro mati c hydr ocar bon s; how ever , they sho uld

not be used with fast burn ing gase s and vap ors, such as hydr oge n, acet ylen e, or olefi nic hydr ocar bon s, with out addi tion al tests .

3.3 Limitations
Whi le liste d arre sters are teste d by labo

rato ries, the test con ditio ns may not be equi vale nt to or repr esen tativ e of the actu al serv ice con ditio ns of parti cula r pipi ng syst em desi gns. The listi ngs indi cate whe ther the test con

ditio ns incl ude d ignit ion at the ope n end of a pipe atta che d to the devi ce. If they indi cate that ignit ion at the ope n end of a pipe was incl ude d, the listi ngs wou ld also indi cate

the max imu m per miss ible leng th of the pipe . This limi tatio n of pipe leng th mea ns that the state d leng th is the grea test for whi ch the arre ster was succ essf ully teste d; this pres ume s

that the pipe is a strai ght secti on of pipe that is the sam e size as the arre ster. To be effe ctiv e, liste d arre sters mus t be inst alle d in acco rdan ce with their liste d inst allat ion para met

ers.

3.4 Test Procedures

The test proc edur e for liste d arre sters whi ch is prov ided in UL 525 defi nes ignit ion as occ urri ng at the ope n end of the disc harg e pipe and und er nonf lowi ng con

ditio ns, so this test proc edur e is not appr opri ate for liste d arre sters in clos ed pipi ng syst ems or with flow ing flam mab le mat erial s. Bef ore arre sters are inst alle d in pipi ng syst ems,

the arre sters sho uld be teste d with a proc edur e that ade quat ely repr esen ts the serv ice con ditio ns und er whi ch they will be used . The para met ers of the test proc edur e mus

t be equi vale nt to the oper atin g con ditio ns, whi ch incl ude fuel mixt ure com posi tion, tem pera ture, pres sure , flow rate, and pote ntial ignit ion loca tion s relat ive to the arre sters . In som e

case s, testi ng und er over driv en con ditio ns or with pres sure pilin g effe cts cons ider ed may be appr opri ate. In othe r case s, testi ng deto nati on arre sters at defl agra tion con ditio

ns sho uld be cons ider ed beca use arre sters suit able for deto nati ons hav e faile d defl agra tion tests [3].

3.5 Untested Arresters

Som e imp rope r inst allat ions of liste d arre sters hav e bee n in

serv ice for year s with out an acci dent , but this sho uld not be cons ider ed as proo f that such inst allat ions are safe. In mos t of thes e case s, the arre sters hav e nev er bee n subj

ecte d to flam e fron ts. The stati stica l risk is low beca use the requ ired mixt ure and an ignit ion sour ce hav e not occ urre d sim ulta neo usly . In facil ities that term inat e at cont inuo

us ignit ion sour ces, such as furn aces , flare pits, or pilot light s, the prob abili ty of a sim ulta neo us occ urre nce of the prop er mixt ure and an ignit ion sour ce is grea ter. Wh ere arre sters

hav e bee n inst alle d in pipi ng bet wee n the gas and ignit ion sour ces, ther e hav e bee n man y inst ance s of flam es occ urri ng with in the pipi ng, mig ratin g thro ugh the

arre sters , and resu lting in an expl osio n.

3.6 Flame Arresters in Series

In man y case s, plac ing two or mor e flam e arre sters in seri es prov ides only slig ht addi tion al prot ecti on whe n com

pare d to a sing le arre ster. If flam e prop agat ion con ditio ns caus e the first arre ster to fail, ther e is a sign ifica nt prob abili ty that an iden tical seco nd arre ster will also fail. In any

case , ther e is little if any test wor k to veri fy the ben efits of arre sters in seri es.

SECTION 4 - UNLISTED ARRESTERS 4.1 Description

Besi des liste d arre sters , ther e are othe r arre stin g devi ces and tech niqu

es in use with in the hydr ocar bon proc essi ng indu stry, incl udin g wat er seal s, pac ked beds , velo citytype arre stin g devi ces, and mec hani cal inter rupt ion of the flam e path . The

se arre stin g devi ces and tech niqu es also hav e limi tatio ns and sho uld be teste d at full scal e to dete rmi ne their effe ctiv enes s und er actu al serv ice con ditio ns.

4.2 Water Seals

Wat

er seal s are ofte n desi gne d and inst alle d to prev ent reve rse gas flow , and their desi gn is pote ntial ly capa ble of prev enti ng flam e prop agat ion [4, 5, 1]. In each wat er seal,

the gas mixt ure is bub bled thro ugh a rese rvoi r of wat er, a proc ess that may prev ent the pass age of flam es. The flam es are inter rupt ed beca use each gas bub ble is isol ated fro m the

next . No stan dard desi gn or listi ng is avai labl e for wat er seal s. Eac h inst allat ion pres ents a spec ific prob lem invo lvin g the rate of the gas flow , the dept h of the seal, and the

size and the conf igur atio n of the vess el that cont ains the wat er. Som e imp orta nt desi gn cons ider atio ns for the wat er seal are as foll ows : a. It should prevent rupturing under flame-produced pressure. b. It should reliably maintain the required water level for normal as well as flame-produced conditions. c. It should protect against freezing.

4.3 Packed Beds

For man y year s, grav el, rasc hig ring s, sma ll peb bles, and othe r bulk mat erial s hav e bee n used as flam e arre sters in pac ked tow ers or colu mns . The re are no esta

blis hed desi gn crite ria for usin g pac ked beds as flam e arre sters .

4.4 Velocity-Type Arresting Devices

Wh ere the flow of a gas mixt ure is limi ted to a sing le dire ctio n, it is poss ible to ensu re, by desi

gn, that the flow velo city will nev er be less than the velo city corr espo ndin g to the max imu m rate of prop agat ion of flam es in the mixt ure und er cons ider atio n [1]. For a sma ll dia

met er pipe disc harg ing gaso line vap ors into the ope n air, an effl ux velo city of 10 feet per seco nd is cons ider ed ade quat e [2, 6]; how ever , the appr opri ate velo city mus t be dete rmi

ned for each case . The appr opri ate velo city can be dete rmi ned fro m the gas mixt ure and pipe dia met er [1]. Con troll ing flow velo city thro ugh a velo citytype arre stin g devi ce sho

uld be rega rded as an effe ctiv e tech niqu e for prev enti ng flas hba cks only whe n the ignit ion sour ce is at the ope n end of the pipe . In the desi gn of a velo citytype arre stin g devi

ce, som e mea ns mus t be prov ided eith er to mai ntai na mini mu m velo city und er all oper atin g con ditio ns or to inter rupt the gas sup ply if the flow velo city bec ome s too low. The

desi gn mus t also prov ide som e mea ns eith er to inter rupt the gas sup ply or to exti ngui sh burn ing with in the velo city secti on of the arre ster. This prev ents the flam es fro m heat ing

the arre ster eno ugh to per mit the m to pass thro ugh it, whi ch can occ ur with in a few min utes.

4.5 Mechanical Interruption of Flame Path

A clos ed pipe valv e can prev ent flam e pass age as long as the valv e

can be clos ed quic kly eno ugh. Usi ng a valv e as a flam e arre ster app ears to be a limi ted opti on, thou gh. Ach ievi ng the rapi d resp onse time for clos ure is diffi cult; how ever , beca

use flam es are acco mpa nied by vibr atio n, pres sure rise, tem pera ture, and ultra viol et emi ssio ns, sens ing devi ces posi tion ed som e dist ance fro m the valv e can be used to clos e it,

prec ludi ng the pass age of flam es. Mec hani cal inter rupt ion is prob ably mor e usef ul in com bina tion with the othe r appr oac hes disc usse d in this publ icati on.

SECTION 5 - SUMMARY
For cell ular and

othe r teste d arre stin g devi ces, the foll owi ng guid elin es are reco mm end ed: a. Only a listed or tested arresting device that is within the range of its listed or tested parameters should be used. b. For applications that use listed arresters outside the range of their test parameters and for all unlisted devices, a fullscale test under conditions equivalent to actual service conditions should be conducted. Actual service conditions include mixture composition, temperature, pressure, flow rate, and ignition location relative to the arrester. If detonation conditions can occur at the arrester, the device must be tested under detonation as well as deflagration conditions. c. The arresting device must be installed and maintained in the exact mechanical form in which it was tested; this includes maintaining the exact form of the element, its housing, and gaskets. d. For piping system applications where both the flammable mixture and an ignition source are likely to be present simultaneously, flame arresters should not be considered as the sole means of protection but should be used as a supplement to other systems or operational controls. e. If the arresting device is not designed to withstand sustained burning on the face, a provision must be made to detect and suppress or prevent burning.

f. When an arrester is installed in a piping system, it should be checked periodically to ensure that it has not been damaged, has not clogged, or has not corroded. Therefore, the inline arrester must be installed in a location that facilitates inspection and required maintenance.

SECTION 6 - REFERENCES
1. Howard, Walter B., "Flame Arresters and Flashback Arresters," Plant/Operations Progress, American Institute of Chemical Engineers, New York, October 1982, Vol. 1, No. 4. 2. Broshchka, G.L., Ginsburgh, I., Mancini, R.A., and Will, R.G., "A Study of Flame Arresters in Piping Systems," Plant/Operations Progress, American Institute of Chemical Engineers, New York, January 1983, Vol. 2, No. 1. 3. Roussakis, N., and Lapp, K., "A Comprehensive Test Method for Inline Flame Arresters," Paper presented at API Committee on Safety and Fire Protection Spring Meeting, Tulsa, Oklahoma, April 13, 1989. 4. Watson, P.B., "Flame Arresters," Paper presented at the Conference on Instrumentation and Safety in the Oil and Natural Gas Industries, Glasgow College of Technology, Glasgow, Scotland, March 9 - 10, 1977. 5. Bartknecht, W., "Mechanical Flame Barriers," Explosions, SpringerVerlag, New York, 1981, pp. 155 - 158. 6. Wilson, R.P., and Crowley, D.P., Performance of Commercially Available Flame Arresters for Butane/Air and Gasoline/Air Mixtures, NTIS No. AD 063002, 1978. Order No. 855-20280 American Petroleum Institute 1220 L Street, Northwest Washington, D.C. 20005