Smoke motion: comparison of experimental data with simulations

Fire safety in buildings usually relies on the withdrawal of smoke or the preservation of stratification to allow proper evacuation of occupants. In order to assess safety levels, much work has been done in the past, leading to regular prescribed solutions. However, in the case of large enclosures such as theatres or sports arenas, the degree of precision of these classical methods is not well known, leading to high calculated safety coefficients and thus high construction and operating costs. This paper is devoted to a specific project where smoke motion simulations were run to help conceive a more cost-effective design. After completion of the building, a large-scale fire test was performed inside to validate the conclusions of the numerical study. The paper focuses on this fire test and subsequent simulations performed on the fire test scenario and aims at comparing the numerical and large-scale experimental results. Several simplifying modelling hypotheses (geometry, fire modelling, boundary and initial conditions) are examined and compared with the test. Provided enough care is taken in the simulations, these results show that good agreement can be reached. Recommendations are drawn on this basis.