CHARACTERISTICS of TunnEl fIRES

Fig. 6: Equivalent heat load for different kinds of vehicles (Putz, 2005)the structure. Fig. 7 shows that the heat-time distribution isalso more advantageous.When designing tunnels, there is no opportunity (with somerare exceptions) to calculate and analyse the individual heatevolving. Instead of describing of fires in terms of a “correct”standard fire, characteristic curves were established to describethe characteristic of an average tunnel fire. The common usedfire-characteristic curves are shown on the Fig. 8.Each curve is based on one or more of the following:various research, suppositions, national standards, flammablematerial compositions and the results of small or large scaletests (Promat, 2006).ISO Standard: The characteristic curve used mostly in thedesign of buildings, which is defined by many EuropeanNational Standards (ISO 834, BS 476:part 20, DIN 4102,AS 1530). It is based on the continuous ignition and naturalburning of flammable materials surrounding the fire. Theequation of the curve is shown in (1). T C 20 345log(8tmin 1)The maximum temperature 0,167 min 2,5 min 20 1080 1 0,325 t can be calculated on tthe basis ofT C e 0,675ethe fire load time defined by the National Standards.The maximum evolving heat mass is around 30 MW. It isnot proposed that the ISO standard “ISO” curve be used in0,167 min 2,5 min 20 1280 1 0,325 t tT C the design of tunnels. eFig. 8 shows 0,675that the estandard “ISO”curve has a relatively long heat accumulation time whichis not acceptable to describe the tunnel fire theory of fastheat accumulation.HydroCarbon: Hydrocarbon fires demonstrate an alternatecharacteristic behaviour. Both the characteristic behaviourFig. 7: Vehicles of small and large combustion heat (Blennemann andand the extreme of the curve differ from the commonlyGirnau, 2005)used standard “ISO” curve. The faster ignition of thehydrocarbons causes a faster increase and larger combustionfires produce higher maximum temperatures. The equationTof the curveC is shown 20 345log(8tin (2).min1)0,167 min 2,5 min 20 1080 1 0,325 t t 0,675T C e eTC 20 345log(8tmin1)0,167 min 2,5 min 20 1280 1 0,325 t t 0,6750,167 min 2,5 min 20 1080 1 0,325 t t 0,675HydroCarbon Modified: Analogue to the “Hydrocarbon” fire,Tthe CFrench Standard defines a efire characteristic ecurve with amaximum T C temperature of 1300 e °C as opposed to ethe 1100 °Cshown above. The equation of the curve is shown in (3).0,167 min 2,5 min 20 1280 1 0,325 t t0,675T C e e (3)(1)(2)Fig. 8: Standardized fire-characteristic curves; air/gas temperatures aroundthe fire (after Blennemann and Girnau, 2005)(Blennemann and Girnau, 2005). However, this evolvingenergy is distributed in time and space. Thereby the total energymass is definable for every kind of vehicle if the componentscan be summated.It is important to note that real fire protection begins withthe applied materials of the vehicles. The total mass of theflammable materials in a vehicle design conforming to modernstandards and considerations is smaller, so the evolving of heatand (toxic) smoke puts less strain on the human organism andBoth the “HydroCarbon” and the “HydroCarbon Modified”fire characteristic curves came from the petrol chemicalindustry into civil engineering. Their use could bedemonstrated because of the huge amounts of benzene andgasoline around the fire, especially in road tunnels.RABT-ZTV: The “RABT-ZTV” fire characteristic curves havebeen determined and standardized in Germany as a result oflarge scale tests. In contrast with the foregoing theories, thecurves are described with break points instead of equations.Different characteristic curves were determined for bothrailway and road tunnels. Both curves rapidly reach themaximum temperature (1200 °C) within 5 minutes. Themaximum temperature is then held for 25 to 55 minutesdepending on the type of curve. The cooling phase is 110minutes for both curves (Table 1).The German fire characteristic curves are used for designas well as for research because of the linear drive. Curveswell define the rapid increase of the temperature in the first58 2008 • CONCRETE STRUCTURES

Table 1: Significant break points of the “RABT-ZTV” curves (Promat, 2006)Fig. 9: Large scale test in Germany (1998) (Haack, 2002)Table 2: Significant point of the RWS, RijksWaterStaat fire characteristiccurve (Promat, 2006)Fig. 10: Maximum air temperatures in the walls during a railway tunnelfire (Richter, 1993)few minutes (Fig. 9). However, this is the only characteristicwhich determines the parameters of the cooling phases.The speed of the cooling is also important, not only for themodel, but also for the residual properties of the material.RWS, RijksWaterStaat: Characteristic curves from theNetherlands determined at 1979. This curve shows thehighest maximum temperature of all the curves above. Itdescribes the worst case scenario on an accident in a roadtunnel when a tanker with 50 m 3 gasoline explodes. As aconsequence of the accident, 300 MJ of energy evolvesin 180 minutes. The behaviour of the curve shows rapidincreasing during the ignition period, deceleration in rateof temperature increase in the second period, a significantmaximum point, slight recession and stagnation (Table 2)Apart from the lack of the cooling phase (the flammablematerial is burn out after 180 minutes) the RWS firecharacteristic curve has the greatest upper bound of almost100% of the tunnel fires.The curves above represent the distribution the maximumdesign value of air or gas temperature in a tunnel versustime.The spatial changes of temperature were also examined. Itwas verified that the maximum temperature develops at theroof of the tunnel; lower temperatures are noticeable at thewalls. Figs. 10-12 shows the distribution of developed airtemperatures during the fire flaming out from different kindof vehicles.The longitudinal distribution also shows the spatial changesextending away from the fire. However by the time of ignition,the secondary ignition of flammable materials around thefire is also possible. Thereafter the maximum points of thetemperature curve are extended over the tunnel’s longitudinalFig. 11: Maximum air temperatures in the walls during an undergroundrailway tunnel fire (Richter, 1993)Fig. 12: Maximum air temperatures in the walls during a road tunnel fire(Richter, 1993)CONCRETE STRUCTURES • 2008 59

Fig. 13: Distribution of air temperature parallel to the longitudinal axis,away from the fire (Richter, 1993)the actual countries. Thereby the special curves are advisablefor the numerical modelling of the effect of fire heat on thewall, as well as for research propose into the structural materialsand the structure itself.This paper demonstrates that the air temperature around theplace of the fire shows a distribution in the cross section aswell as in the longitudinal section of the tunnel. Thereby thetemperature values for the load of the structure are definable.It can be stated that by designing a new tunnel, from whicha burning vehicle cannot escape, the consequences of the heatevolving should be examined. In older tunnels, where heatresistant of structure was not observed, the same considerationsshould be performed and the structure should be equipped withadditional fire protection.Thence in Hungary it is necessary to determine our standardsfor the characterization of tunnel fires. Thereunto researches(full scale test and small models, numerical analysis) are tobe carried out beside the acceptance of European regulationsand standards.5. acknowledgementsAuthor wish to express his gratitude to Mr. Martin Smith andMr. György Posgay for their advices and proof reading of thepresent paper.Fig. 14: Distribution of air temperature parallel to the longitudinal axis,away from the fire (Blennemann és Girnau, 2005)section, and increasingly more tunnel linings will be subjectedto the maximum thermal load as shown on Figs. 13-14.4. ConclusionsDuring a tunnel fire, a huge amount of heat is evolved.Developing heat can only dissipate slowly due to the formof the structure and the surrounding soil and rock. Thereforesignificant values of air temperature can be formed at the placeof fire in the structure.International research provides valuable information aboutthe evolving heat mass based on the theoretical assumptionsin small or large scale tests. Special fire characteristic curveswere drawn up by West European practices which calculatefaster heat accumulation and higher maximum temperaturevalue than the “conventional” characteristics that are valid in6. referencesBeard, A. and Carvel, R. (eds.) (2005): „The Handbook of Tunnel Fire Safety”,Thomas Telford Ltd., LondonBlennemann, F. and Girnau, G. (eds.) (2005): „Brandschutz in Fahrzeugenund Tunneln des ÖPNV”, Alba Fachverlag, DüsseldorfHaack, A. (2002): „Current Safety Issues in Traffic Tunnels”, STUVA, KölnPromat (2006): „Fire Curves”, www.promat-tunnel.comPutz, U. (2005): „Brandbeanspruchung von Tunnelbeton“, 45.Forschungskolloquium des DAfStb, 100. Jahrgang, Beton- undStahlbetonbau, pp. 173-176.Richter, E. (1993): „Heißgasentwicklung bei Tunnelbränden mit StraßenundSchienenfahrzeugen – Vergleich gemessener und in Vorschriftenenthaltener Temparatur-Zeit-Verläufe“, STUVA Tagung ’93 in Hamburg,STUVA, Köln, pp. 131-137.Schlüter, A. (2004): „Baulicher Brandschutz für Tunnelbauwerke: Richtlinien,Vorgaben, die Realität und geeignete Maßnahmen“, Tunnel, November2004 (Sonderdruck: Promat GmbH.)DIN 5510-2 (2003. 09.): Vorbeugender Brandschutz in Schienenfahrzeugen- Teil 2: Brennverhalten und Brandnebenerscheinungen von Werkstoffenund Bauteilen; Klassifizierungen, Anforderungen und PrüfverfahrenSándor Fehérvári (1981), MSc. in civil engineering (BME 2006). PhDstudent at BME, Dep. of Construction Materials and Engineering Geology.Main fields of interest are the effect of the tunnel fires on structures, structuraland curtain grouting, building and repair techniques of underground andtunnel structures. Member of the Hung. Group of fib, the Hung. Chamber ofEngineers, the Hung. Tunnelling Assoc., the Hung. Sci. Assoc. for Transport,the Sci. Society of Silicate Industry (Concrete Div.). Activities: ITA WG 6(Maintenance and Repair).60 2008 • CONCRETE STRUCTURES