- Author
- Soderbom, J.
- Title
- Smoke Spread Experiments in Large Rooms. Experimental Results and Numerical Simulations.
- Coporate
- Swedish National Testing and Research Institute, Boras, Sweden
- Report
- SP REPORT 1992:52, 1992, 31 p.
- Keywords
- smoke spread | experiments | fire models | data analysis | zone models | smoke filling
- Abstract
- SP, fire technology have performed four full scale fire tests in a large industrial building. The experiments were designed to investigate the smoke spread through very large rooms. Pool fires ranging from 7000 to 1800 kW were used. The largest compartment involved in the experiments were the size of a public assembly room, 20 m by 33 m and 2,7 m height. The experimental results have been compared to simulations with a zone model. The experimental results reported are time histories of: Height of the interface between hot and cool smoke gas layer; average temperature in the smoke gas layers; mass burning rate of the fuel; mass flow out of room of fire origin. General observations during the experiments show that: The stratification of smoke gases created a two zone atmosphere with a clear cool lower zone and a hot smoky upper zone. This was established even at the low temperature rises in the lage rooms. In the large rooms, the height of the interface between the hot and cold layer varied with the distance from the fire source. Furthest away from the fire source where the combustion gases were cooler, the interface height came closer to the floor. Even small restrictions in the low velocity flow of, cool combustion gases between compartments cause a break down of the stratified smoke layer. The results from the experiments are important as a set of base data for validation of compartment fire models and also as a guide to the time scale for smoke filling of this type of fire scenario. The data have also been used to validate a smoke spread model. The code FAST version 18 developed at National Institute of Standard and Technology, USA, (NIST) was used. The predictions showed general agreement in so much that the predicted interface height followed the measured values. However, the interface height were predicted with good accuracy only in the room of fire origin. In the room furthest away from the fire room, the two zone stratification of the smoke gas layer showed a break down which could not be predicted by the employed calculation model. The temperature rise was predicted with acceptable accuracy during the first 3 to 4 minutes of the experiments. For this type of scenario, the two zone model, FAST, can be used to estimate the hazard from a fire of the sizes used in the experiments.