On the Characteristics of Fires in Tunnels

Abstract: Important phenomena related to tunnel fires are presented including both experimental and theoretical work. The main focus is on the large-scale fire tests performed in 2003, in the Runehamar tunnel in Norway. During these tests heat release rate (HRR), temperatures, gas concentrations (O2, CO2, CO, HCN, and H2O), velocities, radiation, and smoke were measured. Four tests with a set-up simulating HGV cargos were performed with fuel loads constituting of different mixtures of cellulose and plastic. The fuel load ranged from 2850 kg to 11010 kg giving maximum HRRs between 66 MW and 202 MW. The results showed that ordinary (non-hazardous) cargo can give HRRs and temperatures normally expected to be related only to hazardous goods (e.g. tanker fires). Thus, the highest HRR and gas temperature were far higher than that suggested by most tunnel design guidelines. A close correlation between the maximum HRR and energy content is presented for passenger cars and HGVs. It has also been shown in the Runehamar tests that the fire in an HGV can spread a long distance downstream in a tunnel, of the order of 100 m. Based on the test results and real incidents one must emphasize the importance of the incident operation during the first five to ten minutes of the fire. The actions of the people inside the tunnel and the rescue personnel during this time period can be crucial to the outcome. During two of the fire tests, large pulsations of the gas flow inside the tunnel were observed. These pulsations were registered only when the measured heat release rate was higher than 125 MW - 135 MW. Two different periods of pulsations were registered, short periods of approximately 4 s and longer periods of approximately 18 s. The pulsations are presented and explanations are given, using two different approaches: an acoustic approach and a frequency response analysis based on an impedance approach. The intrinsic resonances of the system proved to be close to the periods found in the experiments. Several factors can affect the pulsations, but the calculations show that the oscillation periods are properties of the system. Other issues discussed include backlayering and the effect of the ventilation on the HRR, flame length, and the conditions in the tunnel. Literature data are, where relevant, compared with the results from the Runehamar tests. In this analysis toxicity and the effect of the ventilation conditions on the gas composition in the tunnel are also included. Special attention was paid to the CO/CO2 ratio and its dependence on the ventilation and combustion conditions. In this context results from a detailed study of the emissions from an enclosure fire under varied ventilation conditions are used as the basis for a discussion of the impact of ventilation on the acute toxicity of the fire gases.