Search

International audience; The aim of this study is to collect data on the combustion of horizontally-oriented poly(methyl methacrylate) (PMMA) samples in reduced oxygen atmospheres for CFD model validation. Experimental results relating the oxygen concentration to the burning behavior of 3-cm-thick clear PMMA slabs are discussed. Experiments are conducted in the controlled atmosphere calorimeter of IRSN called CADUCEE. Pyrolysis and combustion of 0.2×0.2 m2 horizontally-oriented PMMA samples are studied varying the oxygen molar fraction from 0.210 to 0.180, extinction occurring at about 0.175. The measured quantities are the regression rate of the slab, mass loss rate, temperatures and total and radiative heat fluxes at the center of the slab. All experiments are carried out twice, showing a good repeatability. It is found that the slab regression rate, mass loss rate and heat fluxes at the slab center decrease significantly with the oxygen concentration, while the gas temperature is much less sensitive. Most notable is that the radiative and convective contributions to the total heat flux remain almost constant, respectively 0.65 and 0.35. It is also found that both heat fluxes and mass loss rate exhibit linear oxygen-concentration-dependent behavior. From an energy balance and current average values of the total heat flux and regression rate at the center of the slab, the present study obtains a heat of gasification value of 2.25 MJ kg−1, in agreement with literature data. © 2017 Elsevier Ltd

An important cause of fire departure in industrial facilities is due to electrical origin and particularly to electrical cabinets. The investigation of such fires has been scarce up to now and has been investigated exclusively in the nuclear industry. The Institut de Radioprotection et de Sûreté Nucléaire (IRSN) conducted a large number of experiments involving electrical cabinets burning either under a calorimetric hood or inside a mechanically ventilated compartment to investigate this topic. Calorimetric hood experiments demonstrated that the most important parameter is the size of the vents of the cabinet and that the time to flashover depends on many factors and seems somewhat random with regard to the observable parameters. The influence of the compartment on the fire behavior depends on the temperature of the surrounding atmosphere of the cabinet and on the oxygen content in the compartment at the level of the inlet vent of the cabinet. The compartment strongly impacts the pyrolysis of the combustible, affecting the fire duration, but has a weak effect on the Heat Release Rate (HRR). Experiments were usually remarkably reproducible, opening the way to a phenomenological description of this type of fire. A semi-empirical model based on the coupled solution of ventilation limit and excess pyrolysate could then be developed. This model was introduced in a zone code, and an ad-hoc modeling of the fire extinction based on a critical surfacic mass loss rate is proposed. The major features of the compartment fire experiments such as characteristic HRR and fire duration could then be reproduced with acceptable error. The development of such a semi-empirical model is a common practice in fire safety engineering concerned with complex combustibles. © 2011 Elsevier Ltd. All rights reserved.

Experiments were carried out to study the effects of sample width on the early upward flame spread, when typically the heat release rate per unit width ( Q ˙ ′ ) is less than 30 kW/m. Samples of clear poly-methyl-methacrylate (PMMA) were 0.4 m high and 0.03 m thick, and width values of 0.025, 0.050, 0.1, 0.15, and 0.2 m were tested. A new processing method, based on wavelet decomposition, was developed to detect the pyrolysis front location from video recordings of the bubbling surface. Different propagation behaviours are observed depending on the slab width, w. For w⩾0.1 m, flame height, rate of spread (ros), and Q ˙ ′ are found to be width-independent. The flame seems to remain laminar during the entire spreading process and flame height measurements are found to be in good agreement with available literature data. For the narrower slabs (0.025 and 0.05 m), a transition from laminar to turbulent flame is observed, the value of Q ˙ ′ at which the transition occurs decreasing with the width of the slab. Moreover the spread rate decreases with the slab width as a result of reduction in flame height due to three-dimensional aerodynamic effects. For w=0.05 m, flame height decreases whereas Q ˙ ′ is not affected. This may be related to the lateral entrainment of air above the pyrolysis region. For w=0.025 m, three-dimensional effects extend to the burning region. A significant decrease in the heat release rate per unit width is then observed, which in turn leads to a reduction in flame height. This configuration, where the flame tip is very close to the pyrolysis front, is of particular interest for fire safety since it corresponds to near-critical conditions for flame spread.

Large-eddy-simulation (LES) is performed to investigate the transport characteristics and structure of large-scale, turbulent fires on vertical surfaces under natural convection conditions. The combustion process in buoyancy-driven wall fire plumes is assumed to be diffusion controlled, permitting a mixture-fraction-based modeling approach. The three-dimensional, time-dependent Navier–Stokes equations and mixture fraction are solved with sufficient temporal and spatial resolution. The large-scale eddies are simulated directly and subgrid-scale motion is represented by Smagorinsky model. The computed, time-averaged flame height, velocity and temperature profiles are compared with experimental data, and a relatively good agreement is attained. For a given heat release, the air entrainment rate over a vertical wall fire decreases to 1/4 one of a pool fire plume. The predicted entrainment rate closely follows an adjusted entrainment rate correlation from a pool fire plume. It is observed that the flame height behind a pyrolysis region over a vertical wall is more important than that for turbulent jet fire, and also a function of the heat release raised to the nth power with n=2/5.

International audience; Electrical cabinet fires are a major concern in nuclear power plants due to the significant impact on plant safety. This paper investigates electrical cabinet fires in highly confined and mechanically ventilated compartments. In particular, it addresses the issue of experimentally determining the heat release rate of an electrical cabinet fire through realistic large-scale testing. The fire source is highly complex in nature due to the heterogeneity, stochastic geometry and location of the combustible materials within the electrical cabinet. After providing a detailed description of the experiments and apparatus, this paper focuses on the thermal modeling necessary to assess the heat release rate based on the energy balance in the fire compartment, the ventilation system and the electrical cabinet itself. The experimental method is then applied to a propane gas burner and an analytical cabinet containing PMMA tiles with homogeneous spatial distribution in order to validate the thermal model and heat measurements. Finally, the method is used to assess the heat release rate and combustion efficiency of an electrical cabinet. The various terms of the energy balance are described in detail to show the major role of the thermal inertia of both the electrical cabinet and the compartment walls. © 2012 Elsevier Ltd. All rights reserved.

A series of experiments is performed to determine the local burning rate of vertically oriented burning slabs of PMMA (poly(methyl methacrylate)) under steady-state conditions. The burning rate is essential for the determination of flame temperatures, flame lengths, and burnout, all critical parameters to establish fire growth. The experimental procedure consists of a postprocessing of digitized images obtained with three CCD cameras placed on the sides of the PMMA slabs. The cameras are used to establish the displacement of the regressing surface as a function of time. The PMMA/gas interface is characterized by a bubbling layer that scatters light, making the bubbles visible. The video images are processed by establishing a gray-level intensity threshold that defines the depth of the bubbling layer. The local burning rate is then computed from the time derivative of the surface regression. Local burning rates for PMMA slabs, 30 cm high, 3 cm thick, and 2.5, 5, 10, 15, and 20 cm in width are obtained. Three-dimensional effects appear for 2.5-cm-wide slabs. For wider slabs, the local burning rate along the centre line is found not to be width-dependent and to be in good agreement with literature data. Results show that the pyrolyzing surface divides vertically into four regions. In the lowest region (0 < x < 4 cm), the local burning rate increases with the distance from the leading edge, x. For 4 < x < 10 cm, the local burning rate follows a power-law decay. For slabs wider than 2.5 cm, the exponent approximates -0.35. This is higher than that predicted by the two-dimensional laminar boundary layer theory, i.e., -0.25. For 2.5-cm-wide slabs, the local burning rate decreases faster, with a power law dependency of -0.8. From x = 10 to 18 cm, the local burning rate is approximately constant, which is believed to correspond to the laminar/turbulent transition region. In the upper region, the flow becomes fully turbulent and the burning rate increases weakly with x. © 2007 The Combustion Institute.