Brannsikkerhet i naturlig ventilerte parkeringshus

Fire safety in naturally ventilated car parks This study investigates fires in car parks, and is financed by the Norwegian Directorate for CivilProtection (DSB) and the Norwegian Building Authority (DiBK).The main objective of the study is to collect knowledge in order to evaluate whether it is safe to reducethe fire resistance of main load-bearing systems in car parks in fire classes 1 and 2 to R 15 A2-s1,d0[incombustible material], provided that more than 1/3 of the wall area is open and that the buildingdesign is such that good ventilation is ensured. Reduced fire resistance is indicated as a pre-acceptedsolution in the guideline to the Regulations on technical requirements for construction works(TEK17).The results of this study indicate that the fire resistance of the load-bearing structures shouldnot be reduced from R30 - R60 to R15, even if the wall surfaces have more than 1/3 open areafraction.Relevant regulations in Norway, Sweden, Denmark and Finland have been surveyed. The followingmain rules apply to load-bearing systems in car parks in these Nordic countries:• Car park with two floors : R 30 – R 60• Car park with three and four floors : R 60• Car park with more than four floors : R 60 – R 90 – R 120In Norway and Sweden, subject to different prerequisites, the fire resistance may be reduced to R 15.In Denmark and Finland, however, the use of R 15 for the load-bearing system in car parks is notallowed. Sweden and Finland require the installation of an automatic sprinkler system if the fireresistance is reduced. This requirement applies to car parks with two floors in Sweden (from R 30 toR 15 with sprinklers) and for car parks with a height above 28 m, which is around eight floors, (fromR 120 A2-s1,d0 to R 90 A2-s1,d0 with sprinklers). Of the four Nordic countries, only Norway uses theopen area fraction in wall surfaces as a basis for reducing the fire resistance.Under the pre-accepted solutions in the guideline to TEK17, car parks with open wall surfaces will inpractice often need to have sprinkler systems, either because each floor is defined as a separate firecell, or because the total gross area in a fire cell with open connection across several floors exceeds800 m2, or because the fire section size demands it. In these cases, the design will be morecommensurable with Sweden, i.e. a reduction in fire resistance, but with the installation of a sprinklersystem.In order to be able to assess the effect of wall surfaces with a different open area fraction, a total of tenfire simulations with different wind conditions (direction and force) were carried out. Two generic carparks, one with 21 % open area fraction and one with 41 % open area fraction, were examined. Thecar parks have one floor and a floor area of 1 797 m2, three «open» sides and one side closed by afirewall.It is emphasized that the CFD simulations and structural analysis involve a number of uncertaintiesand limitations. Absolute values for fire spread and collapse time therefore only provide some2© RISE Research Institutes of Swedenindications and no final answers. The focus is therefore on a comparison of car parks with a differentopen area fraction.In a car park with 41 % open area fraction, i.e. more than 1/3 open wall area, the main load-bearingsystem may under certain conditions be constructed with a fire resistance of minimum R 15 A2-s1,d0[incombustible material]. For a car park with 21 % open area fraction, the fire resistance must beminimum R 30 or R 60 depending on the number of floors.In this study, only one floor was simulated. The fire simulations are based on a simple spreadingmodel and are well suited for a comparative study. Owing to model uncertainty, spreading progresscannot be directly used for the analysis of other car parks.The fire simulations have shown that a larger open area fraction, and thus better ventilation, can limitthe extent of fire spread, i.e. the number of cars to which the fire spreads.The difference between a closed (less than 1/3 open area fraction) and an open (more than 1/3 openarea fraction) car park in terms of the number of cars that are burning, is most visible after 40 minutes– 60 minutes, when the fire has reached a certain size. This is because the difference in the time ittakes the fire to spread between cars is accumulated over time, and because the ventilation conditionsassume greater importance when the fire becomes so large as to make it ventilation controlled.For very high wind velocities (e.g. 11 m/s), the open area fraction plays a smaller role, since this leadsto good ventilation also when the open area fraction is lower (21 % in this study).Increased ventilation and thus increased wind velocity in car parks, leads to the fire spreading faster inthe wind direction, downstream of cars that are already burning. This is because the fire and smoke aredriven to one side, and thus closer to adjacent cars. In major fires, increased ventilation will also givethe fire increased access to oxygen.A faster spread of fire in the wind direction may result in more cars burning simultaneously, comparedwith a more closed car park, where the flow rate is lower. Several cars burning simultaneously maycause greater thermal stress on the support system, and potentially an earlier collapse of the structure.The extent of stress will depend significantly on the wind direction, layout of the car park, location ofthe fire start relative to the location of other cars, and so on.A simplified structural analysis showed that an increased open area fraction both entails a positive anda negative effect on the structure’s load-bearing ability in a fire, depending on whether windconditions are favorable or not.Regardless of wind conditions, the structural analysis showed that expanding the open area fractionfrom 21 % (i.e. less than 1/3) to 41 % (more than 1/3), has a smaller effect on the collapse time thanreducing the fire resistance from R 30 to R 15. The difference is even more pronounced in a reductionfrom R 60 to R 15. By using R 60 none of the beams collapsed. The results of this study indicatetherefore that the fire resistance for load-bearing structures should not be reduced even if the wallsurfaces have more than 1/3 open area fraction.For all the fire simulations visibility conditions were examined after 15 minutes. For very low or veryhigh wind velocities little difference in visibility conditions is expected, depending on the open areafraction. At moderate wind velocities, statistically the most common, it turned out In what way open wall surfaces impact a car park fire is highly dependent on the fire scenario andwind conditions. These two factors cannot be controlled. Dimensioning the fire resistance to the mainload-bearing system in a car park based on the open area fraction of wall surfaces (more than 1/3 ofthe area) is therefore considered unreliable. Open wall surfaces contribute in some cases to improvingvisibility conditions in car parks, which may extend the available escape time. For this reason, openwall constructions are nevertheless considered advantageous.This study did not examine the effect of sprinkler systems in combination with a reduction in fireresistance, such as is allowed in Sweden. Nevertheless, a sprinkler system, which is little affected bywind conditions, is generally considered better suited as a compensatory measure if the fire resistanceis reduced.It is, therefore, our recommendation that the possibility of reducing the fire resistance in open carparks in fire classes 1 and 2 be reconsidered. This option should be considered removed, or othercriteria could be employed to reduce the fire resistance, such as e.g. sprinkler systems (as in Sweden).Sprinkler systems are considered better suited as a compensatory measure if the fire resistance isreduced.As a basis for such reassessment experiments (fire tests) should be carried out. This is becauseCFD simulations have some limitations, especially regarding the interaction between sprinkler/dropsof water and solid fuel.In addition to the wall design, other factors may also impact the spread of fire, such as e.g. ceilingheight, the design of the floor slab, and the distance between cars. These factors were not examined inthis study. E.g., simulations show that a ribbed floor slab (floor slab with underlaying beams) mayhave a large impact on local flow conditions and thus the spread of fire. To examine these parametersit is recommended that the existing fire spreading model be used, and if relevant validated throughexperimental research.Dette prosjektet er finansiert av Direktoratet for samfunnssikkerhet og beredskap (DSB) ogDirektoratet for byggkvalitet (DiBK), og er utført som en del av prosjektporteføljen underforskningsavtalen mellom DSB og RISE Fire Research. Prosjektets bakgrunn og målsetting erbeskrevet i kapittel 1.