Detection of fires in the toilet compartment and driver sleeping compartment of buses and coaches : Installation considerations based on full scale tests

(1)

Detection

of

ﬁres

in

the

toilet

compartment

and

driver

sleeping

compartment

of

buses

and

coaches

—Installation

considerations

based

on

full

scale

tests

Ola

Willstrand

*

,

Jonas

Brandt,

Robert

Svensson

SPTechnicalResearchInstituteofSweden,FireResearch,Box857,SE-50115Borås,Sweden

ARTICLE INFO

Articlehistory: Received3August2015

Receivedinrevisedform13November2015 Accepted18November2015

Availableonline2December2015

Keywords: Firedetection Buses

Toiletcompartments Sleepingcompartments Fullscaletests

ABSTRACT

Effectivefiredetectionsystemsproperlyinstalledinbusandcoachtoiletcompartments and driver sleeping compartments may save human lives and property loss. Rapid detectionallowsforearlyevacuationandextinguishmentofasmallfire,whilelateorno detectionmayallowthefiretospread.Thepurposeoftheworkpresentedinthispaperisto providerecommendationsonhowtoinstallfiredetectionsystemsintoiletcompartments and driver sleeping compartments. The recommendations also cover what type of detectionsystemismostsuited.Asabasisfortherecommendations,fullscalefiretests wereperformedwithdifferentdetectionsystems.Thefiretestswereconductedinrealistic mockupsofatoiletcompartmentandasleepingcompartment.Differentheatandsmoke detection systems wereanalyzed atdifferent positions fordifferentfire scenarios to provideinformationonhowtobestinstalldetectionsystemsinthesecompartments.Five differentscenarioswererunandthemostinterestingfindingwasthattworealisticfire scenariosinthetoiletcompartmentdidnotactivatefiredetectorsintheceilingatrealistic airflowrates.Itisveryrarethatfiredetectorsareplacedanywhereelsethanontheceiling intoiletcompartmentsonbusesandthefirewouldthenbeverylargeupondetection.

ã2015TheAuthors.PublishedbyElsevierLtd. ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/

licenses/by-nc-nd/4.0/).

1.Introduction

Firesinbusesandcoachesareverycommonandonaverageseveralbuses1_worldwide_are_involved_in_a_ﬁre_incident_each

day.Forinstance,intheUSapproximately160busfireswerereportedeachyearbetween2004and2008[1].InAustralia thereareabout70busfiresperyearresultingininsuranceclaims[2]andinSweden,Norway,andFinlandaboutonepercent ofallbusesinservice,willsufferfromafireincidenteachyear[3,4].

Ifpassengershavereducedmobilitytheevacuationtimemaybeseverelyextended.Forinstance,20elderlypeoplediedin abus_fireinHannover2008[5].The_firewascausedbyashortcircuitinanelectricalcablenearthetoiletandspreadviathe toiletcompartmenttothepassengercompartment.Withaneffectivefiredetectionsystemthistragedymighthavebeen prevented.However,notallfireincidentsleadtofatalities,butthepropertylossandthecostduetorescueoperation,traffic jam,andcleanupcanbeextensive.Theenvironmentaleffectsofboththefireitselfandextinguishingagentsmayalsobe

*Correspondingauthor.Tel.:+46105165000,Direct:+46105165450;fax:+4633135502. E-mailaddress:ola.willstrand@sp.se(O.Willstrand).

1_In_this_paper_the_term_“bus”_refers_to_buses_as_well_as_coaches. http://dx.doi.org/10.1016/j.csfs.2015.11.002

2214-398X/ã2015TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/ licenses/by-nc-nd/4.0/).

ContentslistsavailableatScienceDirect

Case

Studies

in

Fire

Safety

(2)

severe[6].Alltheseeffectsmightbemitigatedwithaneffectiveﬁredetectionsystemthatenablesearlyevacuationand suppressionoftheﬁre.

Basedonreported_fireincidentsinbusesandcoachesthe_firesmostfrequentlyoriginateintheenginecompartmentorin thewheelwell[1,4].Severalstudiesoffireprotectionintheenginecompartmenthaverecentlybeenperformed[7–9],and ongoingfiredetectionprojectsforthesespacesarerunninge.g.,atSPTechnicalResearchInstituteofSweden.However,this paperfocuson_firedetectioninthetoiletcompartmentanddriversleepingcompartmentandnoextensivestudyon_fire detectioninthesecompartmentshas,toourknowledge,beenreported.Inmedia,recentbusfireshavebeenreportedwhere thefirestartedinthetoiletcompartment[10,11]andinthecatastrophicbusfireinHannover2008thefirealsostartedinthe interiorofthebus[5].ThestudyreportedinthispaperwaspartlytriggeredbythenewUNECErequirement,regardingfire detectioninthetoiletcompartmentanddriversleepingcompartmentofbuses,thatcameintoeffectinJuly2014[12].The newrequirementstatesthatanexcesstemperatureorsmokeshallbedetectedinthesecompartments.

Reportedinthispaperisaninvestigationofwhattypesofdetectionsystemsaremostsuitableinthetoiletcompartment anddriversleepingcompartmentofbusesandhowtobestinstallthesystemsinthesetypesofcompartments.Themain questionsansweredbythispaperarehowdifferenttypesofdetectionsystemsplacedatdifferentpositionsrespondto differentﬁresandhowtheventilationconditionsmayinﬂuencetheresponsetime.

2.Methoddescription

Firedetectorsystemsweretestedatdifferentpositionsinrealistictestmockupsofthetoiletcompartmentandthedriver sleepingcompartmentofbuses.Differentﬁresourceswerepositionedatdifferentlocationsinsidethemockupsandtests wereperformedunderdifferentventilationconditions.

2.1.Mockups

Statisticsonheight,width,anddepthfortoiletcompartmentsanddriversleepingcompartmentswerecollectedfor 26differentbusesandmeanvalueswereusedforthedesignofthemockups[13].ThemockupsareshowninFigs.1and2. Thetoiletcompartmentsofbuseshaveingeneralquitesimilardimensionsandthelargestdifferenceswerefoundbetween toiletcompartmentspositionedintherearofthebusandtoiletcompartmentsindouble-deckerscomparedwithtoilet compartmentspositionedin thestaircase,which isthemostcommonlocation.The dimensionsofthedriversleeping compartmentswerefoundtovarymorethanfortoiletcompartments,butalsoforthesethemeanvaluesofthesurveywere used.Thedepthwasinmostcasesthewidthofthebus.Thesleepingcompartmentmockuphasdecreasedceilingheightin themiddlesectionwhichisduetothegangwayinthepassengercompartment.Thisdecreaseisnotexistentinallbusesand itcouldalsovaryinsize,butitwasincludedinthemockupbecauseitdelayssmokedistributionwhichseverelyaffects detectorsnotplacedinthedirectvicinityoftheﬁre.

(3)

Theventilationconditionsintoiletcompartmentsandsleepingcompartmentswereexaminedbymeasurementsintwo differentbusesandbyobtaininginformationfromaWCsystemmanufacturer.Thebusmanufacturerswerenotableto provideinformationabouttheairflowthroughthetoiletcompartmentandsleepingcompartment,butonlytheairflow entering the passenger and driver compartment. There are differences between different buses, but there are also similarities.Intoiletcompartmentsairissuckedoutviaaseparatefanandexitsunderthebus,whichappliestoallbustoilet compartmentssincethisconfigurationpreventsodorsfromreachingotherareasofthebus.Theairentersviagapsaround thedoorinmosttoiletcompartments.However,somecompartments,especiallyreartoiletcompartments,dohaveafeed fromtheairconditioningsystem.Inthemockuptheairinletwaslocatedintheupperrightcornerofthetoiletcompartment door,similartosomebuseswhichhavealargergapatthedoorhinge.Theairoutletoccurredthroughthreedifferentholes leadingtoanenclosurecontainingthefan.Twooftheholeswereairvents(mostbuseshaveoneortwoofthese)andone holewastheopeningforthetrashcan.Notallbuseshavethisconfigurationofthetrashcan,theremightbeotherlocations andsometimesatoponthetrashcan,butthiswasthoughtasthemostinterestingcasesincefiresmokefromthetrashcan mightbepreventedbytheairflowfromreachingthetoiletcompartment,withimportantimplicationsfordetection.

In thedriver sleepingcompartment theventilationconditions differmore betweenbuses,but generallythere isa manually operated fan and some passive inlet/outletchannels. In the mockup a fan was placed in one end of the compartment suckingoutairand intheotherendthereweretwoairinlet vents.Thiswasthoughtastheworstcase conditionsincetheairﬂowmaydelaythesmokespreadtotheairinletsectionofthecompartment,increasingtheresponse timeofadetectorplacedinthatsection.Fanpositionsandairinlets/outletsaremarkedinFigs.1and2.

2.2.Testscenarios

DuringthetestsdifferentventilationconditionswereusedwithapproximateairflowsinaccordancewithTable1.Inthe toiletcompartmentthehighfanspeedconfigurationisthemostcommoncondition,buttheairflowdependsmuchonhow sealedthetoiletcompartmentissincethefannormallyhasafreeblowingcapacityof160–220m3_/h._For_an_untight_toilet

compartmenttheairflowcouldthenbemuchhigher.Sometoiletcompartmentshavetwofanmodes,whenthetoilet compartmentisunoccupieditrunsatalowerspeedthanifthecompartmentisoccupied.Thelowfanspeedcon_figurationin thetestsisprimarilyarealisticairflowforthesetoiletcompartments.

Forthedriversleepingcompartmentthetestswereperformedwithahighfanspeedmodeandwiththefanturnedoff.

Fig.2. Mockupofthedriversleepingcompartmentinbuses,seenfromtwooppositesides.

Table1

Differentfanconﬁgurationsusedinthetests.

Highfanspeed Lowfanspeed

Airﬂowthroughtoiletcompartment 60–90m3_/h _20–30_m3_/h

(4)

ThetestsperformedarepresentedinTable2.Eachtestwasruntwicewithverysmalldifferencesindetectiontimesand ﬁredevelopment.Temperatures,obscurationmeasurements,anddetectiontimespresentedareaverages.

InthecigarettetestastandardcigarettefromNIST(NationalInstituteofStandardsandTechnology)wassmokedand consumedinaboutoneminute.Inthepapertestsatrashcanfullofpaperhandtowelswasignitedbyahotwire.Thesizeof theheptanepoolwas1010cm2_and_it_was_positioned_on_the_ﬂoor_in_the_large_toilet_compartment_space._The_heptane_pool

isnotarealisticfiresourceinthetoiletcompartment,butwasusedbecauseofgoodrepeatabilitycomparedtotheotherfire sources.Theplasticsandrubberfiresourcewaspositionedunderthesink,intheconcealedspacewherethefanwaslocated, symbolizingapump,cables,andotherelectronicdevicescontainedhere.Itwasignitedbyahotwireasforthepapertests. Themattress,polyetherfoamwithcottoncover,inthesleepingcompartmentwasignitedbyahotwirethroughthecornerof themattress.Notestwasrunlongerthanfourminutesandinsomecasesonlyforabouttwominutes.Thefirespeakedor stabilizedinthistimeandthetestswerecancelledwhennomorefirealarmswouldhavechangedtheresultsorconclusions. Temperaturesandsmokeobscurationcurvesforallfirescanbefoundinthetestreport[13].

2.3.Detectors

ThedifferenttypesofdetectorsusedinthetestsareshowninTable3.Allsmokedetectorsarebasedonthephotoelectric principle.Thereasonwhynoionizationsmokedetectorsweretestedisduetothefactthatthephotoelectricdetectorsare morecommonamongthevehicleﬁrealarmandsuppressionsystemsuppliers.Onereasonforthatisincurredregulatory costformanufacturing,transport,anddisposalofionizationsmokedetectors[14].Inaddition,theaimofthetestingwasnot tocomparetheresponsetimeofdifferentphotoelectricorionizationsmokedetectors,butrathertocomparedifferent detectorplacementsand tocomparepointsmokedetectorstoaspiratingsystemsandheat detectors.Thesmoke/heat detectorshaveonesmokesensorandoneheatsensorthatactivatedseparatelyinthesetests.Alldetectorsarecommercial approveddetectorsanddifferencesinsensitivityreﬂectmarket.Theactivationlevelsoftheaspiratingsystemsarestatedfor themeasuringchamberandifthedetectorsamplesairfrommorethanoneplacethesmokewillbedilutedbeforeitreaches themeasuringchamber.Smokeobscurationmeasuredindecibelpermeter(dB/m)isthereductionoflightpassingthrough thesmoke.

Thetestedaspiratingsmoke/heatdetectorsusestandardpointsmoke/heatdetectorstogetherwithasamplingsystem, whichmakesitlessexpensivethantraditionalaspiratingsmokedetectorsystemsthataresensitiveenoughtousemany samplingholes.Inthetestseachofthistypeofdetectorsampledairfromoneposition,whereasthemoresensitiveaspirating smokedetectorsampledairfromtwopositions.Noteintheresultshownumberofsamplingholesandsensitivityaffectsthe detectiontimesofsmokedetectors.

Intheteststhedetectors,thermocouples(TC),andobscurationmetershaddifferentpositionsandthesepositionsare listedandexplainedinTables4and5andFig.3.EachnumberinTable5representsaseparatedetector,exceptforthe aspiratingsmokedetectorwhichusestwosamplingpointsindicatedbyaplussign.Theobscurationmeterinthetoilet compartmentceilingcoversbothposition1and2.

Table2 Testscenarios.

Test Firesource Fireposition Ventilationcondition

1 Cigarette Toiletcompartment,seatlevel Lowfanspeed

2 Paper Toiletcompartment,trashcan Lowfanspeed

3 Paper Toiletcompartment,trashcan Highfanspeed

4 Heptanepool Toiletcompartment,ﬂoorlevel Lowfanspeed

5 Heptanepool Toiletcompartment,ﬂoorlevel Highfanspeed

6 Plastics/rubber Toiletcompartment,abovefan Lowfanspeed

7 Plastics/rubber Toiletcompartment,abovefan Highfanspeed

8 Mattress Sleepingcompartment,fansection Highfanspeed

9 Mattress Sleepingcompartment,fansection Nofan

Table3

Thefourdifferentdetectorsusedinthetests.

Activation

Pointsmokedetector 0.5–1.0dB/m

Pointsmoke/heatdetector 0.1–0.15dB/m,54C Aspiratingsmoke/heatdetector 0.1–0.15dB/m,54_C

(5)

3.Results

InTable6theresponsetimesofalldetectorsinthedifferenttestsaresummarizedwhilefurthercommentsconcerning

thetestsaregiveninthefollowingsections. 3.1.Cigarettetest

Theonlydetectorinitiatinganalarminthistestwastheaspiratingsmokedetector,whichalsoisexpectedtobethemost sensitiveoneaccordingtoTable3(eventhoughtwosamplingholeswereused).Theaspiratingsmoke/heatdetectors(pos. 2and4)gavepresmokealarms,whichmeansthatthesmokeobscurationalarmlevelwasreachedbutsincethesmoke concentrationthereafterdecreasednoconfirming firealarmswereinitiated.Accordingtothesmokedetectorsandthe obscurationmetersitwasapproximatelythesameamountofsmokeattheceilingasunderthesink,butahigherfanspeed wouldprobablyhavereducedthesmokeconcentrationintheceiling,sincethesmokewouldbedrawndirectlytowardsthe fan.Nocigarettetestwithhighfanspeedwasperformedhowever.Thatmostofthedetectorsdidnotgiveafirealarmon cigarettesmokeisatleastpartlyduetothefactthatthesedetectorsaredesignedtohaveahighresistancetofalsealarms.In thisscenariothatmeansthattheobscurationduetothecigarettesmokewastoolow.Asimplerdetectormaybemore sensitivetocigarettesmoke,butcouldalsobemorepronetofalsealarmsduetoe.g.,dust.

3.2.Papertests

Alldetectorsintheconcealedspaceunderthesinkwereactivatedinthepapertests,butnoneofthedetectorslocatedat theceilingactivated.Thesmokedetectorsactivatedalreadybeforetherewerevisibleflames.Themaindifferencebetween the low fan speed and thehigh fan speed test was that withthe low air flowsome smokeentered themain toilet compartmentspaceafterawhile,whichwasseenbytheobscurationmeterandthethermocouplesattheceiling,butforthe highairflownosmokeorheatenteredthemaintoiletcompartmentspace.Fig.4showshowtheflamesareheldbelowthe trashcanopeningbytheairflow.

3.3.Heptanepooltests

Intheheptanepool teststhedifferencebetweenthelowfanspeed andhighfan speedtestwasverysignificant.A comparisonofthetemperaturegraphsinFigs.5and6 showsthatthesequenceisalmostreversedsuchthatthepositionof highesttemperaturewithlowairflowalmostbecamethepositionoflowesttemperaturewithhighairflow.Thefirewas positionedontheairinletsideandwithlowairflowthetemperaturewashighestonthissideoftheceiling.Athighairflow mostofthesmokewassuckedoutviatheairventsandtrashcanholebeforereachingtheceilinginthetoiletcompartment. Thedetectorsattheceilingwereabout20sfasterthanthedetectorsunderthesinkwiththelowfanspeedconfiguration,but

Table4

Explanationofpositions.

Toiletcompartment

1 Ceilingleft(fan/trashcanside) 2 Ceilingright(airinlet) 3 Attheopeningofthetrashcan

4 Undersink(intheconcealedspacecontainingthefan)

Sleepingcompartment

5 Ceilingfansection

6 Ceilingmiddlesection 7 Ceilingairinletsection 8 Ceilingaboveﬁreorigin

9 Wallabovefan(half-heighttotheceiling)

Table5

Positions of detectors, thermocouples and obscuration meters in the mockups.

Positions Pointsmokedetector 1,2,4,5,7,9 Pointsmoke/heatdetector 1,2,5,6,7,8 Aspiratingsmoke/heatdetector 2,4,6 Aspiratingsmokedetector 1+4,5+9 Thermocouples(TC) 1,2,3,4,5,6,7,8

(6)

withthehighairflowitwasaboutthesameresponsetimesforalldetectorsexceptthosepositionedontheairinletsideof theceilingwhichwerenowabout20sslowerinresponsetime.Notableisthatthehigherairflowdoesnotaffectthe aspiratingdetectorsasmuchasthepointsmokedetectors.Thereasonthatthethermocoupleunderthesinkstartedsome degreeshigherthantheotherswasthatthefiretestswererunwithshortintervalandsomeheatremainedintheceilingof thissmallconcealedspace.

3.4.Plastics/rubbertests

Theresultsoftheplastics/rubbertestsweresimilartotheresultsofthepapertestsinthetrashcan,whichmeansthat withhighairflownoheatorsmokeatallenteredthemaintoiletcompartmentspacewhilewithlowairflowthesmoke brokethroughtheair-barrier.Theonlydifferencewasthatthesmokeproductionwassomuchhigherthaninthepaperfire thatalsothedetectorsintheceilingwereactivatedquitefastinthelowfanspeedtest.

Fig.3. Drawingonpositions.Firepositionsaremarkedforpaper(trashcan)andplastic/rubberfireintheconcealedspace,heptanepoolfireonthefloorand cigarettethroughthewallinthetoiletcompartmentandthemattressfireinthesleepingcompartment.

(7)

3.5.Mattresstests

Inthemattressteststheﬁrewaspositionedinthefansectionofthesleepingcompartment,suchthatthesmokehadto moveagainsttheairﬂowtoreachtheairinletsection.Themaingoalofthistestwastoseethetimedifferencebetween detectioninthefansectionandintheairinletsectionwithandwithouttheimpactofafan.Contrarytoexpectations detectionintheairinletsectionwasfacilitatedbythefan,duetothefactthatthefancausedcirculationinsidethesleeping compartment.Thetimedifferencebetweendetectioninthefansectionandtheairinletsectionincreasedfromabout10sto about20swhenturningoffthefan.

Themattressﬁresourcewasanalyzedfurtherregardingtoxicelementsinthefumes,seerightsideinFig.7.Themattress was ignitedunderthehoodoftheconecalorimeter[15](withtheconicalheaterremoved)andthetoxicfumeswere analyzedwithaFTIR-spectrometer(FourierTransformInfraredSpectroscopy).Highlevelsofcarbondioxide(CO2),carbon

monoxide (CO), hydrogen cyanide (HCN), and nitric oxide (NO) were detected from the mattress. As expected, the concentrationoftoxicelementsinthefumesfollowedthesmokeobscurationcurve,whichmeansthattheymayberelated totheobscurationmeasurementsinthesleepingcompartmentmockup.Theshort-termexposurelimitssetoutbythe occupationalhealthauthorityinSweden(“Arbetsmiljöverket”)[16],i.e.,acceptablelevelsfor15minexposure,werereached atabout0.5–3dB/msmokeobscurationforthemattressﬁresource.Thisisthepointwheremostsmokedetectorsinitiatean alarm(includingentrydelayandprocessingtimeofthedetector).At10dB/msmokeobscuration,reachedinthetestsafter 1.5–2minfromtheignition,highlevelsoftoxicsubstancesweremeasured:about5%CO2,800ppmCO,70ppmHCN,and

250ppmNO.Thisisabout5–8timeshigherthantheshort-termexposurelimitsandaccordingtotheNationalInstitutefor Occupational Safety and Health(NIOSH) theselevels areimmediately dangeroustolife and health. TheirlistedIDLH (ImmediatelyDangeroustoLifeandHealth)valuesofthementionedsubstancesare4%ofCO2,1200ppmofCO,50ppmof

HCN,and100ppmofNO[17].Theresponsetimesofthedetectorsinthesetestswerearound60s,whichdonotgivemuch timeleftforevacuation.

Table6

(8)

4.Discussionandconclusions

AccordingtothenewUNECEReg.no.107requirementsmokeorheatdetectorsshallbeinstalledintoiletcompartments anddriversleepingcompartmentsofbuses.Theperformedtestshaveresultedinvaluableinformationofwhattoconsider wheninstallingthesedetectors.Smokedetectorsaregenerallymuchfasterthanheatdetectors,whichisalsoconfirmedin alltestsperformedinthisstudy.Intheteststhefireshavedevelopedquiterapidly,butforslowgrowingfiresthebenefitof smokedetectors compared to heat detectors willbe even greater. There are locations where heat detection maybe considered,e.g.,intheconcealedspaceunderthesinkintoiletcompartmentsorclosetothetrashcanwherethedetectoris expectedtobeintheimmediatevicinityofthefire.Inboththepapertestandtheplastic/rubbertestthetemperatureinthe concealedspaceunderthesinkwasover100Cafteroneminute.Inverynarrowspacesandinothercircumstanceswhenthe detectorisclosetothepotentialfiresourceheatdetectorswillalsoreactrelativelyquickly,althoughsmokedetectorswill mostoftenstillbefaster.Thebenefitsofusingheatdetectorsinthesespacesarethattheyareusuallycheaperandmore robust.Theymayalsorequirelessmaintenanceandinspectionsthansmokedetectorsthatmustbeinspectedregularlyto ensureproperfunctioning.However,heatdetectorsshouldonlybeusedasacomplementtosmokedetectors.

Intoiletcompartmentsitiscommontoinstallasmokedetectorintheceiling,butthetestsclearlyshowsthatwithan operatingfanitcouldbedifficulttodetectatrashcanfireorcablefiresolelywithadetectorintheceiling.However,thefan maybemalfunctioningresultinginthesmokebeingtransportedupwardsandnotintotheconcealedspace.Insuchcasea detectorintheconcealedspacewouldbeoflimitedusewhileadetectorintheceilingwouldbemoreeffective.Theremight alsobeotherfirescenariosthanthosetestedinthiswork.Thereforeadetectorintheceilingisusefulasapartofanintegral detectorsystem.Thisstudysuggestthatthedetectionsysteminbustoiletcompartmentsshouldconsistofatleastasmoke detectorintheceilingandheatorsmokedetectorintheconcealedspaceofthefan,especiallyifthisspacealsocontainsthe

Fig.4.Paperﬁreinthetrashcan,seenfrominsidethetoiletcompartment.

(9)

trashcan.Twodetectionpointslikethisisverycommonintoiletcompartmentsofairplanes[18].IntheHannoverbusfire with20casualties[5]thefirestartedinacablesomewherebeneaththetoiletandeventhoughafiredetector,regardlessof position,wouldhavebeenafirststepofimprovement,thepositioncouldhavebeencriticalforachievingenoughtimefor evacuation.

Ifsmokedetectorsareusedinmanyspacestheuseofaspiratingsystemscouldbeconsideredinsteadofpointsmoke detectors.Thebeneﬁtofthisapproachisthatonlyonedetectorisneededandthesystemsamplesairfrome.g.,boththe ceilingandotherspacesinthetoiletcompartment.Moreadvancedaspiratingsystemscouldpotentiallyalsosampleairfrom differentlocationsaroundtheentirebus.Anaspiratingsmokedetectorinthetoiletcompartmentceilingalsohasagreat advantageinthatthedetectorishiddenandprotected.Accordingtothebusoperatorstheyhaveproblemswithpassengers pullingdownthedetectors.

Anotherimportantdesignconsiderationwheninstallingdetectorsinthetoiletcompartmentceilingistheneedtoavoid theairflowfromtheairinlet.Thetestshaveshownthatthedetectiontimemaybedelayedconsiderably,andthedelaytime maybeevenlargerforslowgrowingfires.Notableisthatthehigherairflowdoesnotaffecttheaspiratingdetectorsasmuch asthepointsmokedetectors.

Inthesleepingcompartmentteststheresponsetimedifferencebetweendifferentdetectorplacementswasrelatively small,whichindicatesthatonedetectormaybesatisfactory.However,forrapiddetectiontheuseoftwosmokedetectors shouldbeconsideredifthedecreasedceilingheightinthemiddlesectionisconsiderable.

Theconclusionsabovearebasedonthepresentedtests,whichincludeassumptionsanddesignparameterswithadegree ofuncertaintyandvariability.Eachfirescenariowasruntwiceandthetemperaturedeviationscouldforsomefirescenarios bemorethan50_C,_but_for_stable_fires_as_the_heptane_pool_fire_it_was_not_more_than₁₀_C._However,_in_all_tests_the

sequencesinwhichthedetectorsactivatedandifthedetectordidnotactivatedwerethesameforbothtestsanditisthis informationthatisusedfortheconclusions.Ifthedesignparametersarechangedtheresultcouldbedifferent,butthetests were performed such that the conclusions should be relevant for most toilet compartments. Regarding important parametersasairﬂowandlocationofairinletandoutletthetestsarefocusedonthemostinterestingconﬁgurationsofthose

Fig.6.Temperaturesofthermocouples(TC)inTest5:heptanepoolﬁre,highfanspeed.

(10)

thatarecommon.However,wheninstallingdetectorsincompartmentsverydifferentfromtheconﬁgurationsofthesetests, itisimportantthatthecompartmentisfurtheranalyzedwiththeconclusionsinthispaperinmind.

Acknowledgment

Thisworkwas_ﬁnancedbytheFFIprogramoftheSwedishGovernmentalAgencyforInnovationSystems,VINNOVA. References

[1]N.R.Meltzer,G.Ayres,M.Truong,Motorcoachﬁresafetyanalysis:thecauses,frequency,andseverityofmotorcoachﬁresintheUnitedStates,Firesin Vehicles—FIVE2012,Chicago,USA,2012.

[2]TheOfﬁceofTransportSafetyInvestigations(OTSI),BusSafetyInvestigationReport—AnInvestigationintoBusFiresinNSW2005–2012,2013. [3]R.Hammarström,J.Axelsson,M.Försth,P.Johansson,B.Sundström,BusFireSafety,SPTechnicalResearchInstituteofSweden,SPReport2008,41,

2008.

[4]E.Kokki,Busﬁresin2010–2011inFinland,FiresinVehicles—FIVE2012,Chicago,USA,2012.

[5]R.Damm,TheHannoverbusfireandactivitiesonimprovingfiresafetyinbuses,FiresinVehicles—FIVE2010,Gothenburg,Sweden,2010. [6]P.Blomqvist,Emissionsfromfires—consequencesforhumansafetyandtheenvironment,DoctoralThesis,LundUniversity,Lund,2005. [7]J.Brandt,M.Försth,Testingactivefireprotectionsystemsforenginecompartmentsinbusesandcoaches—apilotstudy,SPTechnincalResearch

InstituteofSweden,SPReport2011,22,2011.

[8]P.Smith,A.Chattaway,J.Peoples,Acomparisonofvariousﬁredetectionmethodologiesintransitvehicleenginecompartmentﬁreprotectionsystems, SAEInt.J.Commer.Veh.5(1)(2012)343–353,doi:http://dx.doi.org/10.4271/2012-01-0983.

[9]R.Ochoterena,M.Hjohlman,M.Försth,DetectionofFiresintheEngineCompartmentofHeavyDutyVehicles,ATheoreticalStudy,SAETechnicalPaper 2014-01-0423,doi:10.4271/2014-01-0423,2014.

[10]Busﬁreduetotoiletruling,[Online].Available:<http://www.dailyexpress.com.my/news.cfm?NewsID=88185/>(accessed17.04.14).

[11]R.Seales,Miracleescapefordozensofschoolchildrenastheyfleeblazingbusaftertoiletcaughtfire,[Online].Available:<http://www.dailymail.co.uk/ news/article-2107120/Miracle-escape-dozens-school-children-flee-blazing-bus-toilet-caught-fire.html/>(accessed17.04.14).

[12]UNECERegulationNo.107—Rev.3—Amend.4,Generalconstructionofbusesandcoaches,[Online].Available:<http://www.unece.org/ﬁleadmin/ DAM/trans/main/wp29/wp29regs/R107r3am4e.pdf/>2012.

[13]O.Willstrand,J.Brandt,R.Svensson,Firedetection&ﬁrealarmsystemsinheavydutyvehicles,WP5–Firedetectioninbusandcoachtoilet compartmentsanddriversleepingcompartments,SPReport2014,28,2014.

[14]FireIndustryAssociation(FIA),IonistationChamberSmokeDetectors(ICSD):Applicableregulationsformanufacturing,transportanddisposal,FIA FactFileNo.0017,2005.

[15]TheInternationalOrganizationforStandardization(ISO),ISO5660-1,Reaction-to-ﬁretests—Heatrelease,smokeproductionandmasslossrate—Part 1:Heatreleaserate(conecalorimetermethod),2002.

[16]Arbetsmiljöverket,Kemiskaarbetsmiljörisker,AFS2011:19,ISBN978-91-7930-560-4,ISSN1650-3163,2011.

[17]NationalInstituteforOccupationalSafetyandHealth(NIOSH),DocumentationforImmediatelyDangerousToLifeorHealthConcentrations(IDLHs), 1994.[Online].Available:<http://www.cdc.gov/niosh/idlh/intridl4html/>(accessed2.04.14).