Sprinkler i vegtunneler

NVF-seminar Tunnelsäkerhet ur trafikantens synsvinkel Stockholm 20-21 oktober 2010 Sprinkler i vegtunneler Ragnar Wighus Sjefsforsker SINTEF NBL Norges branntekniske laboratorium http:\nbl.sintef.no\ R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 1

Spørsmål som behandles: Brannutvikling i tunneler Ulike typer av aktive brannbekjempelse: Sprinkler Deluge Vanntåke Skumanlegg =vannbaserte brannbekjempelsesanlegg Hvordan kan sprinkleranlegg virke i vegtunneler? Virker ventilasjon og aktive brannbekjempelsesanlegg alltid positivt på personsikkerheten? Hvordan bør aktive brannbekjempelsesanlegg testes ut og dokumenteres? R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 2

Noen referanseprosjekter: UPgrading of existing TUNnels Project No: GRD1-2001-40739 UPTUN WP 2, Task 2-4, innovative technologies Storskala tester av vanntåkeanlegg i 100 m lang testtunnel Hobøl, Norge R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 3

Overdekket brann, 15-20 MW R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 4

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Four pools burning,20mw Shortly after ignition Just before system activation Just after system activation Some time after system activation Outside view after system activation R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 5

Runehamar testtunnel, Åndalsnes, Norge: 1520 m lang, 9 m bred, 6 m høy, 1-3% stigning, gammelt E80-profil Fullskala tester av brannutvikling med ventilasjon, tunge kjøretøy med Last (Europaller, Dieselolje) Fullskala tester av tunnel-kledning av ulike typer (Duk, sprøytebetong med ulike tilslag) Fullskala tester av slokkeanlegg (Skum, vanntåke) Eies og driftes av Vegdirektoratet. Branntester utføres av SINTEF NBL as R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 6

Runehamar 2003 UPTUN SP, SINTEF Tungtransport, Europaller Runehamar 2005 SINTEF (RWS) Europaller, Dieselpool 2,5 m/s ventilasjon Skumanlegg Krav: Etter 15 minutter Radiation kw/m² Air temperature ºC 20 Metres upstream the border of the fire load 5 50 30 Metres upstream the border of the fire load 3 50 0,5 Metres downstream the border of the fire load 12,5 280 R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 7

Brannutvikling i tunneler En personbil utvikler ca 5 MW branneffekt. Varighet: 10-20 minutter Lokalt vil det oppstå temperaturer 800-1100ºC, men disse vil kunne senkes ved ventilasjon En mindre lastebil, buss utvikler ca 20 MW I større deler av tunnelen vil det oppstå temperaturer 800-1100ºC, men disse vil ikke påvirkes så mye ved ventilasjon. Varighet: 20-30 minutter R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 8

Brannutvikling i tunneler En fullastet trailer med brennbare varer utvikler 150 200 MW. Varighet: 1-2 timer med høy intensitet, 3-4-timer med redusert intensitet. Video fra forsøk i Runehamartunnelen R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 9

Hvordan kan sprinkleranlegg virke i vegtunneler? Vann som bringes inn i brannsonen reduserer temperaturen i forbrenningsproduktene (røyk, varme gasser) Dette kan redusere ødeleggelse av tunnelen og infrastruktur i tunnelen Vann demper brannutviklingen dersom det kommer inn i forbrenningssonen Vann kjøler omgivelsene, for eksempel kjøretøyer og last i nærheten av brannsonen og hindrer brannspredning R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 10

Hvordan kan brannbekjempelsesanlegg virke i vegtunneler? Aktivering av dysene kan være avhengig av ventilasjon Branner i tunneler er ofte skjulte, inne i og under kjøretøyer Vanndråper ovenfra er ikke det mest effektive Tradisjonelt sprinkleranlegg er ikke ansett å være optimalt (ventilasjonsfølsom utløsning av enkeltdyser, mye vann) Med et riktig dimensjonert vannbasert brannbekjempelsesanlegg reduseres branneffekten til 30-50% ved branner på 15-20MW R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 11

Konsekvens - frekvens Det har oppstått og vil oppstå tilløp til katastrofebranner i tunnel En alvorlig tunnelbrann kan medføre: Tap av mange liv Store skader på tunnel og infrasturktur Langvarig avbrudd Ved SINTEF NBL undersøker vi konsekvensene av brann og virkningen av bekjempelsesanlegg R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 12

Virker ventilasjon og aktive brannbekjempelsesanlegg alltid positivt på personsikkerheten? Dersom ventilasjon og vannbaserte brannbekjempelsesanlegg er testet sammen og tilpasset hverandre vil det være en positiv effekt ved at branneffekten reduseres Ved mindre branner vil ventilasjon kunne fortynne og transportere ut forbrenningsproduktene Ved større branner kan ekstra luft føre til mer trekk og gi større branneffekt. I slike tilfeller er brannen katastrofal uansett Mer vann og mer luft tilført vil føre til bedre forhold for overlevelse i store branner R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 13

Hvordan bør aktive brannbekjempelsesanlegg testes ut og dokumenteres? Full skala: Samvirke mellom brann, ventilasjon bekjempelsesanlegg og tunnel-geometri (inkludert kjøretøy). Sjiktning, backlayering, sikt, deteksjon, varsling, rømningsforhold, redningsinnsats, Stor skala: Effektivitet av bekjempelsesanlegg ved ulike brannstørrelser, ved ulike kjøretøy, lokal belastning på tunnelkonstruksjon Laboratorieskala: Materialprøvning (toksisitet, brennbarhet og brannmotstand av tunnelkledning), parametervariasjon for ulike konsepter CFD-beregning : Sensitiviteststudier, simulering av redningsinnsats (krav at beregningsmodellene er verifisert) R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 14

Advarsel: For liten skala og for liten realisme i testscenariene kan føre til underestimering av risikovurdering ved tunnelbrann R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 15

Konklusjoner Brannbekjempelsessytemer: Er mest aktuelt i høytrafikk tunneler Innovative systemer bruker ny teknologi, vanntåke framfor konvensjonell sprinkler (mindre vann, aktivering basert på deteksjon av flamme, evt video) Vanskelig å slokke branner i kjøretøyer. Hindre spredning er det viktigste, samt å sikre overlevelse i tunnelen SINTEF NBL as 16

Konklusjoner Vannbaserte brannbekjempelsessystemer vil redusere branneffekten (MW) og senke temperaturene i tunnelen vesentlig Dette vil minske skadene på tunnel og infrastruktur til lokale skader Forholdene oppstrøms brannen vil forbedres vesentlig, redningsinnsats vil være mulig med vinden i ryggen Overlevelsespotensialet for mennesker oppstrøms vil øke vesentlig, og i noen grad også nedstrøms brannen. SINTEF NBL as 17

Takk til: Vegdirektoratet Norge v. Harald Buvik som har utviklet og holder vedlike Runehamartunnelen og som satser på fullskala verifikasjon av brannbeskyttelse i tunneler Rws (Reichswaterstaat) Nederland som har satset på fullskala testing av brannbeskyttelsessystemer SP (Svenska Provningsanstalten) som i en årrekke har fremmet forståelsen for sikring mot tunnelbrann, blant annet ved internasjonale konferanser ComputIt, Trondheim, som har utviklet Kameleon FireEx (KFX ), en CFD-modell med styrke innen brannutvikling, brannlast og virkning av vanndråper SINTEF NBL as 18

Vedlegg til presentasjonen: Definisjoner Noen diagrammer og figurer som underbygger foredraget En presentasjon fra konferansen: International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 19

Noen definisjoner: Sprinkleranlegg: CEN 12845: Sprinkler (automatic): Nozzle with a thermally sentitve sealing device which opens to discharge water for fire fighting Sprinkler (open): Sprinkler not sealed by a thermally sentitve element Delugeanlegg: Flere åpne dyser som åpnes samtidig ved aktivsering av en ventil R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 20

Dysetyper R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 21

Vanntåke: FM 5560 Water Mist Nozzle A special purpose device containing one or more orifices designed to produce and deliver an atomized water spray meeting the definition of Water Mist or meeting the specific requirements of an FM Approved water mist fire test protocol. Nozzles can be designed to operate independently of other nozzles, as a group of nozzles or a combination of the two. R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 22

FM 5560 Deluge System A water mist system using open nozzles attached to a piping system that is connected to a water supply through a valve that is opened by means of a detection system installed in the same area as the mist nozzles. When the valve opens, water flows into the piping system and discharges through all nozzles attached to the system. R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 23

Skumanlegg: Vann med tilsetning av et skumdannende middel som utløses ved brann. Skummet vil danne et sjikt som begrenser avdampningen fra væskeoverflater og andre brennbare flater og begrenser varmestrålingen mot disse flatene. R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 24

Skjebnen til ulike dråpestørrelser R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 25

Conclusions from UPTUN The following conclusions may be drawn on the basis of the tests with the low and high pressure water mist systems: The efficiency of the water mists systems was satisfactorily for both systems. However, the efficiency was strongly dependent on the size of the fire (or heat generation rate), nozzle type, location and the water discharge rates. For the smallest fires (less than or equal to 5 MW) the mitigation effect was minor The best results were achieved for the largest fires (i.e. a heat release rate at or above 20 MW). The mitigation of the heat release rate was down to 20 % of the free burning rate, i.e. a maximum reduction of the heat release rate of 80 %. A rapid reduction of the temperatures downstream from the fire was noticed after activation of the suppression system. The efficiency of both water mist systems was satisfactorily with respect to both heat stresses and the toxicity of the fire atmosphere on human beings. The visibility was not improved downstream of the fire during the first minutes after activation of the suppression systems, but the visibility was generally increased as the fire size and the heat release rate were reduced during fire suppression. The problem of backlayering (i.e. smoke spread upstream of the fire) and the visibility upstream were also significantly improved after activation of the systems. High pressure water mist are using less water and suppresses fires more in the gas phase of, while low pressure is more cooling the fuel surfaces R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 26

Pool-brann, 100 m2 dieselolje, 200 MW R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 27

Pallestabel, opp mot 200 MW R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 28

Stabel med 720 paller, opp mot 200 MW R.Wighus NVF-seminar Stockholm 20-21 oktober 2010 SINTEF NBL as 29

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany The UPTUN project: Original plan for testing Innovative Systems Experience with two water mist systems Conclusions Challenges to innovative water-based systems SINTEF NBL as 30

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany UTPTUN has a strong focus on fire development and fire mitigation: Mitigation measures will be evaluated from current technology, and innovative cost efficient tools will be adapted and verified for tunnel use. The effects of mitigation on tunnel fires shall be evaluated for different ventilation regimes in terms of corresponding changes in heat release rate, temperatures, toxicity and visibility. Especially the hazards associated with Heavy Goods Vehicles (HGVs) need consideration in different tunnel environments since for these vehicles the uncertainty in associated design fire scenario is the largest. Fire propagation to neighbouring vehicles will be specifically considered since it has played a dominant role in the damage levels observed in previous incidents. All this is done to achieve safe and optimised cost efficient mitigation as function of design scenario and tunnel environment. SINTEF NBL as 31

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Fire development - WATMIST MODEL (SINTEF NBL as) Recirculated fire products Fuel release rate Fire products Air exchange Complete mixing Heat and mass conservation Reduced burning rate due to stoichiometry Extinguishment due to oxygen/temperature criterion SINTEF NBL as 32

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany SINTEF report NBL F06115 June 2006 Data Report: High Pressure Water Mist System: Full scale testing of fire mitigaiton in well ventilated tunnel fires Kristen Opstad and Are W. Brandt SINTEF NBL as 33

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany The fire suppression system tested is a high pressure system developed by Fogtec and Semco. The system has all nozzles located in three pipelines in the ceiling of the tunnel. The water pressure is provided by a high pressure water pump located outside the tunnel (a mobile unit). Typical values for the suppression system: Water flow rates from 140 l/min 550 l/min, (2.3 kg/s 9.2 kg/s) Cover area from 230 m 2 290 m 2 Discharge rates from 0.5 l/[min m 2 ] 2.3 [l/min m 2 Water pressure : 60 bar to 120 bar Number of nozzles used from 7 to 14. SINTEF NBL as 34

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany 56 fire tests were performed in a 100 m long tunnel, made from concrete with a cross-section of ~40m 2. The tunnel was rented from IF-Sikkerhetssenter, Hobøl Norway.(South of Oslo) Number of tests Type of fire test 3 One single pool 17 Two pools 22 Three pools 11 Four pools 3 Solid fuel 8 Reference tests without suppression SINTEF NBL as 35

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Side view of the test tunnel SINTEF NBL as 36

5.05m 5.05m 5.9m 5.9m ø 1m International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany End view of the tunnel openings INLET GATE OUTLET GATE 4.5m 8.1m 4.5m 8.1m Some restrictions are made in the cross section at the ends to represent a longer tunnel. Expansion of the gases will in long tunnels lead to high velocities, and hence pressure loss. SINTEF NBL as 37

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany SINTEF NBL as 38

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Partly covered fire, three pools SINTEF NBL as 40

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Solid fuel arrangement, with wood pallets SINTEF NBL as 41

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Measurements: Water supply characteristics (Pressure, flow) Velocity profile downstream fire Temperatures in many positions Visibility (Optical density) Heat fluxes by Plate Thermometers Gas concentrations (Oxygen,CO, Water vapour) Heat release rate (HRR) based on Oxygen depletion SINTEF NBL as 43

Test No Fire load Maximum HRR (MW) Time to reach max HRR Time (after CAF activation) to reduce temperatures downstream of fire * below 280 C 6.3.1 480 pallets, 20% plastic 6.3.2 480 pallets, 20% plastic 6.4.1 720 pallets, 20% plastic 6.4.2 720 pallets, 20% plastic 5.1 100 m 2 Diesel pool 90 (70) 7 min 8 min 35 sec 80 (80) 3 min 13 min 150 (170) 3-4 min 17 min 160 (160) 3-5 min 6 min 120 (150) 1-2 min 2 min 15 sec SINTEF NBL as 47

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Conclusions A total of 64 successful tests are conducted, where 8 tests are reference tests. The heat release rate (HRR) is ranging from about 4MW to about 25MW. Four different pool fire scenarios are included combined with two ventilation conditions. The effect of the mitigation system is strongly dependent on nozzle type and the water discharge rate. For some fires, the mitigation effect was not measurable, while for the best test results, the mitigation of the heat release rate was down to 40 % of free burning rate. The smallest 5 MW fires were hardly affected by the suppression system and the best results were achieved for a 15 MW fire. Expected mitigation for fires between 10 MW to 20 MW is about 30 % 50 % compared to free burning rate. SINTEF NBL as 50

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Conclusions (continued) For these fires, well ventilated conditions applies, the toxicity follows proportional to the heat release rate. The mitigation of toxicity is the same as the reduction of heat release. For the visibility, no clear conclusion can be taken. Normally the visibility is reduced within the first minute of activation of the mitigation system, but it improves as the fires decrease in size. In general, these fires give very low visibility and it is assumed that it is not possible to walk or escape if one is trapped in the smoke. Smoke stratification is hardly observed downstream of the fire in these tests, both in cases with or without mitigation. For smoke spread upstream of the fire, smoke stratification was observed. SINTEF NBL as 51

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Conclusions (continued) Based on toxicity and temperature measurement, it is still possible to survive hours for most of these fires after efficient mitigation. It gives fire brigades time to fight the fire and to rescue personnel captured in the smoke. Available time will depend on the fire size, ventilation condition, cross section area of the tunnel and health of the person escaping. This will not be further evaluated in this report, but will be analysed in other UPTUN reports. SINTEF NBL as 52

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Challenges to Innovative Systems: Integration of firefighting system with ventilation Optimum location of nozzles with regard to fire and obstructions Optimum water application Well fitted detection and zone-based activation system Integration with possible compartmentation techniques SINTEF NBL as 53

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Acknowledgements: This paper presents the fire mitigation program conducted in the EUproject UPTUN. Acknowledgement is given to European Commission (Research Directorates) and the entire 41 partner participants in UPTUN (www.uptun.net) SINTEF NBL as 54

International Tunnel Safety Conference by IWMA and ITA COSUF 2-3 April 2008 Münich Germany Nisser og dverge bygger i berget Men vi skal mine dem alle herut Thi mens vi synger muntre i klynger Sprenger vi berget i stykker med krutt Gremlins and gnomes build inside the mountain But we will mine them all out When we are singing, merry in groups We blow up the mountain into pieces with gunpowder Zipfeln und Zwerge im Bergen bauen Aber wir sollten alle heraus sprengen Während wir singen munter im Haufen Sprengen wir die Berge im Stücken mit Schieß-pulver SINTEF NBL as 55