The problem of fires in tunnels is reviewed and new experimental data are presented which show for possibly the first time that the established correlations between fire size and critical ventilation velocity appear to be incorrect at large fire sizes; the dependence of these data upon the particular tunnel geometry is also evident. The departure from previous expectations is attributed to non-Boussinesq effects and tilting of the fire plume in the downstream direction. The Boussinesq approximation, implicit in standard correlations, is considered to be inappropriate for a study of the dynamics of hot smoke layers due to the significant effects of large density gradients on the horizontal momentum equation.

Aircraft structures are designed to withstand further to dynamic mechanical loadings thermal loads too. In the event of an external fire (while the aircraft is on the ground) the fuselage structure has to withstand and delay fire penetration. Prolonged burnthrough time is the design target. In the current work, a progressive fire-degradation material model is developed that links decomposition kinetics with the thermophysical properties of polymer composite materials. The material model is then implemented in a FE model to simulate the response of the flat panels under fire burnthrough conditions. Experimental investigation is performed in accordance to the ISO2685:1998 (E) Standard.

No abstract

The behaviour of the fire at London's King's Cross Underground Station on November 18, 1987, was the subject of a prolonged and detailed scientific investigation by the Health and Safety Executive on behalf of the Official Enquiry. In the initial stages of the investigation, a detailed damage survey was made at the site of the fire, and samples of many different materials were collected and subjected to a variety of ignition and flame spread tests. In the course of the investigation, fire growth studies were made on a fulil-size section of escalator, on scale models of the escalator, shaft and ticket hall, and by numerical modelling.

No abstract