The Dalmarnock Tests comprise a set of fire experiments conducted in a real high-rise building in July 2006. The two main tests took place in identical flats, Test One allowing the fire to develop freely to post-flashover conditions while Test Two incorporated sensorinformed ventilation management. The test compartments were furnished with regular living room/office items and fully instrumented with high sensor densities. The furniture and objects acting as fuel were arranged to provide conditions that favour repeatability. A full description of the set up of the tests, including fire monitoring sensors, is provided. Focus is on the larger Test One fire for which the major events are reported together with a thorough characterisation of the fire using sensor information. The main aim of the experiments was to collect a comprehensive set of data from a realistic fire scenario that had a resolution compatible with the output of field models. The characterisation of Test One provides a platform with potential for analytical and computational fire model validation.

No abstract

A review of research into the burning behavior of large pool fires and fuel spill fires is presented. The features which distinguish such fires from smaller pool fires are mainly associated with the fire dynamics at low source Froude numbers and the radiative interaction with the fire source. In hydrocarbon fires, higher soot levels at increased diameters result in radiation blockage effects around the perimeter of large fire plumes; this yields lower emissive powers and a drastic reduction in the radiative loss fraction; whilst there are simplifying factors with these phenomena, arising from the fact that soot yield can saturate, there are other complications deriving from the intermittency of the behavior, with luminous regions of efficient combustion appearing randomly in the outer surface of the fire according the turbulent fluctuations in the fire plume. Knowledge of the fluid flow instabilities, which lead to the formation of large eddies, is also key to understanding the behavior of large-scale fires. Here modeling tools can be effectively exploited in order to investigate the fluid flow phenomena, including RANS- and LES-based computational fluid dynamics codes. The latter are well-suited to representation of the turbulent motions, but a number of challenges remain with their practical application. Massively-parallel computational resources are likely to be necessary in order to be able to adequately address the complex coupled phenomena to the level of detail that is necessary.

The Heat Release Rate (HRR) is a critical parameter to characterise a fire. Different methods have been developed to estimate it. The most widespread techniques are based on mass balance. If the heat of combustion of the fuel is known, the measure of the mass loss allows its evaluation. If the burning material can not be identified, calorimetric principles can be used. They rely on oxygen consumption (OC) or carbon dioxide and carbon monoxide generation (CDG) measurements. Their asset comes from the observation that the amount of energy release per unit mass of O 2 consumed or per unit mass of CO 2 produced is relatively constant for a large number of materials. Thus, an accurate HRR can be obtained without knowing the composition of the burning fuel. The aim of this work is to assess this last statement and define how essential the knowledge of the chemistry to calculate HRR for complex materials such as polymers including fire retardants and/or nanocomposites, energetic materials or pine needles is. This assessment ends in an OC and CDG calorimetry comparison of several materials in order to investigate the propensity to determine whether converging or diverging HRR results when average energy constants are used.

No abstract