Numerical Simulation of a Metro Train Fire
Degree GrantorUniversity of Canterbury
Degree NameMaster of Engineering
This research thesis presents the simulation of fire growth and flame spread within a metro train in an underground trainway using Fire Dynamics Simulator (FDS) Computational Fluid Dynamics (CFD) model. The motivation of the study is to predict the heat release rate (HRR) and specifically the peak value for emergency tunnel ventilation system design. Even though currently there are several methods that can be used to estimate the HRR for a metro train, it appears that the current methods cannot realistically predict the HRR because factors such as the burning behaviour of materials; and/or the train and tunnel geometries that affect the HRR are not considered. This project attempts to incorporate these factors in the FDS model. The study evaluated the design of metro trains proposed for the new Circle Line under construction in Singapore. In this research, two modelling approaches were proposed. The first modelling approach prescribed the Cone Calorimeter heat release rate per unit area (HRRPUA) while the other prescribed the heat of vaporisation. The difference between the two is the prescribed constant which governed the rate of pyrolysis. Cone Calorimeter tests were conducted for the surface exposed materials to evaluate the train car materials’ reaction to fire and to derive the material properties for input into the FDS model. In this research, three common fire scenarios have been identified for simulation. They were fire on top of the seat (arson), fire in the corner (arson and electrical fault) and undercarriage fire (electrical fault). The common fire scenarios were expanded to account for ventilation factors. A total of 13 credible fire scenarios were investigated. As it has been found that prescribing the material properties values derived from the Cone Calorimeter test data were not able to accurately predict the ignition and fire growth for the materials simulated. A combination of derived and calibrated properties values has been prescribed in the final simulations. II In the final simulations, the two modelling approaches predicted the same fire severity for different fire scenarios. The simulations indicated that it is important to prevent direct airflow through the train compartment as it may support fire spread if there is a large ignition source. The simulations also indicated that for a scenario that will progress to flashover under the influence of high ventilation airflow velocity, the fire spread to adjacent cars will be very rapid if there is no door installed between the metro train cars. Two peak HRR values have been proposed for the design of emergency tunnel ventilation system for the metro train under consideration based on the simulations. A peak HRR value of 5 MW has been proposed for a metro train fire at the station trackway and a peak HRR value of 10 MW has been proposed for a metro train fire in the tunnel.