A study of repeatability of hot smoke test in a subway station

Hot smoke test is a key method for fire studies in metro systems. But most hot smoke test scenarios are conducted only once, which ignores the experiment uncertainties and errors. To quantify the uncertainty and repeatability of hot smoke tests in subway stations, seven experiments are repeated on a station platform. 120 thermistors are deployed among the platform to acquire the temperature curves in the experimental area. The concept of statistical analysis and functional analysis is adopted to study the error and turbulence caused by random effects. The results indicate that: (1) According to the fluctuation properties, the experiment area is divided into fire plume area and ceiling jet area. The data uncertainty of the fire plume area is significantly higher than the ceiling jet area. (2) The repeatability of temperature data in the fire plume area is poor. The difference between single experiment curve and average curve might reach 23%-24%. (3) The repeatability of the data in the ceiling jet area is much better with a deviation smaller than 7.8%. The validated experiment data presented in this work could be used for simulation model verification. The repeatability analysis method presented is applicable for other fire studies.