This paper aims to characterize dynamics of a fire in the Large-Scale Demonstrator Malveira Fire Test, a full-scale fire experiment carried out in a disused industrial building in Portugal. The Malveira Fire Test is the second stage in the series of full-scale experimental programmes developed for the Real Fires for the Safe Design of Tall Buildings project at the University of Edinburgh. This experiment is intended to act as a real-building demonstration of fire dynamics in large open-floor plan compartments and has as objective to provide a data set to contrast methodologies aiming at design fire inputs representative of real fire dynamics in compartments typical of tall buildings. The Malveira Fire Test showed three distinct fire behaviour modes characterised by the ratio between the velocities of the fire front (") and the burnout front (#$

Knowledge of the first principles defining fire behaviour in large enclosures remains limited despite their common use in modern tall buildings. The evolution of a fire in large enclosures can be defined by the relationship between the flame front and burnout velocities (𝑉 " /𝑉 $% ). The mechanisms governing flame spread and burnout are investigated using four full-scale enclosure fire experiments with high porosity wood cribs with similar enclosure geometries. Flame and burnout fronts position and velocity are estimated using video data. Velocities are affected by the heat feedback from the enclosure and smoke layer to the fuel. The spread velocity shows two regimes, a critical heat flux below which there is no spread (𝑞̇( ,*+,-.. ) and a heat flux that defines the onset of very rapid flame spread (𝑞̇+ (,*+,-.. ). A phenomenological model is developed to help identify the underlying mechanisms controlling the transition between the different spread modes. Both the model and data show that 𝑞̇( ,*+,-.. is controlled by the fuel's surface temperature ahead of the flame front, and that 𝑞̇( ,*+,-.. reduces as the surface temperature approximates steady state. The magnitude of 𝑞̇+ (,*+,-.. is constant and is mainly delivered by the flame heat flux. The dependence of the burnout front velocity to the external radiation is found to be weak. NomenclatureA area b stick thickness C crib burning constant hT total heat transfer coefficient Lv latent heat of vaporisation 𝑚 . mass per unit length 𝑚̇0 .. burning rate per unit area 𝑚̇1 .. free burning rate per unit area 𝑚̇+ .. radiation enhanced burning rate per unit area 𝑞2 ..

Our understanding of fire behaviour and heating conditions for load-bearing structural elements was developed from an immense body of research in small under-ventilated compartment fires. Within the context of contemporary architecture, large open-plan compartments are commonplace, yet understanding of the first principles that define fire behaviour in such enclosures remains limited. Past experiments have revealed that fires in open-plan compartments exhibit three distinct fire spread modes: a fully-developed fire, a growing fire, and a travelling fire. This paper studies the thermal characteristics arising from these fire spread modes and the effects of the ventilation imposed. An experimental analysis of the energy distribution and spatial heating is conducted on a series of large-scale compartment fire tests, with the fire spread mode and ventilation conditions systematically varied. Each fire spread mode is shown to induce significant and characteristic spatial heat distributions. Moreover, the analysis of the ventilation modes shows equivalent thermal loads imposed on the structure in cases where the opening areas are large, and plume flows are dominant despite lower gas temperatures and irradiation. Thus, fires in open-plan compartments pose unique and possibly more severe thermal loading to structural systems, a characteristic not captured by current design fire methodologies.

No abstract

No abstract