Effects of ventilation on evacuation in tunnel fires
MetadataShow full item record
Norwegian road traffic, including transport of goods, is growing and so has the need and ability to build long and deep tunnels. A tunnel fire has a catastrophic potential, so technical measures including ventilation systems are used as replacement for emergency exits or additional tunnel tubes which are very expensive. Originally, tunnel ventilation systems were implemented to handle waste gases and dust from the tunnel, but later this has become one of the most important mitigating measures during fire. The most commonly used ventilation system in Norway are longitudinal ventilation driven by jet fans in the ceiling and using tunnel portals as supply and exhaust. The reason for this is mainly the cost and the fact that transverse systems with vertical shafts are unwanted in densely populated areas, unpractical for tunnels deep into rock and not feasible for sub-sea tunnels. The transverse systems with ducts increase the required tunnel cross-section and become overly expensive. Longitudinal ventilation systems that achieve critical velocity are able to create a smoke-free environment upstream the fire, but often evacuees get trapped at the downstream side of the fire. The issue of downstream evacuation has raised discussions in the tunnel fire safety community and the industry regarding ventilation strategies. This has led to the use of several ventilation strategies, ranging from the ventilation system not starting at all before evacuation is completed to systems running at full power from start. The arguments for not starting the ventilation system is that this can diminish smoke stratification whilst on the other side, arguments are based on the dilution of smoke to an extent that visibility and toxicity does not severely affect the evacuation. This thesis investigate the effects of various ventilation velocities on downstream evacuation conditions using fire- and egress simulation models FDS and STEPS. Fire input data of the simulations are based on full scale tunnel fire test data, including the effect of forced ventilation on the fire growth rate. The egress simulations provide data which is used to calculate FED values on an individual level when combined with the fire simulation results, including the effect of reduced visibility on the walking speed. The results are unambiguous during the important initial phase of evacuation; increased air velocity worsen the downstream conditions severely compared to lower or no ventilation. This is based on measurements in 2 m height downstream the fire for visibility, temperature and CO concentrations. Egress calculations including upstream evacuation show that low air velocities delay the backlayering and downstream smoke propagation sufficiently to allow for safe evacuation at both sides of the fire.The maximum CO concentration was found to be higher for low air velocities, including a no-ventilation scenario. However, for the first 10-14 min at 200 m distance downstream of a relatively severe HGV fire, the highest air velocities caused significantly higher CO concentrations at 2 m height. Simulations with the same fire growth rate for high and low air velocities further confirm this. At 600 m downstream, the difference between high and low air velocities increases. Here, the CO concentration for lower air velocities remain well below the higher air velocities for the duration of the simulations. The somewhat surprising results, at least for some, are believed to be caused by two main phenomena's which should be investigated closer, potentially, through full scale fire tests so that more explicit recommendations can be given for ventilation strategies; A. Forced air movement, induced from jet fans placed in the ceiling, push smoke towards the floor at an increasing rate with increased velocity. This is both a result of the location of the jet fans and the dilution of smoke that cools it down and makes it descend faster. The friction between the smoke and air increase with increased velocity and disperse the smoke so that it is both diluted and that the direction is changed from a uniform horizontal flow to include vertical directions as well, but in smaller scale. B. Increased ventilation velocity spreads the smoke faster downstream the tunnel. Even if it does not obtain the same maximum CO levels as the smoke from lower ventilation velocities, visibility is reduced at an earlier stage and exposure time is increased for evacuees.