Bernard Porterie

A formal averaging procedure for radiation heat transfer in particulate media

Abstract Radiative heat transfer in participating particulate media is modeled using a formal volume averaging procedure. The multiphase medium is composed of emitting–absorbing–scattering phases, i.e., a gas phase and several particle phases. Each particle phase contains large, opaque, gray, diffuse, and spherical particles having locally the same geometrical, thermophysical, and radiative properties. The resulting multiphase radiative transfer equation (MRTE) is solved using the discrete ordinates method. The present computed results are found to be in good agreement with those obtained using the Monte-Carlo theory and with the available experimental results. The coupling effect of the MRTE with the averaged energy equations in a three-dimensional cavity which is differentially heated or which contains a volumetric heat source is studied. A parametric study is performed for particle-phase and gas properties, and wall emissivity.

Fire spread across pine needle fuel beds: characterization of temperature and velocity distributions within the fire plume

The aim of this article is twofold. First, it concerns the improvement of knowledge on the fundamental physical mechanisms that control the propagation of forest fires. To proceed, an experimental apparatus was designed to study, in laboratory conditions, the flame of a fire spreading across a pine needle fuel bed. Characterization of temperature was managed by using a reconstruction method based on a double thermocouple probe technique developed recently. The vertical gas velocity distribution was derived from the previous reconstructed signals by measuring the transit time of a thermal fluctuation between two points of the flow. Second, the experimental data were used for the testing of a physical two-phase model of forest fire behavior in which the decomposition of solid fuel constituting a forest fuel bed as well as the multiple interactions with the gas phase are represented.

Predicting wildland fire behavior and emissions using a fine-scale physical model

ABSTRACT A physical fine-scale two-phase model has been developed for the purpose of determining wildland fire behavior and emissions. The situation modeled corresponds to a spreading wildfire driven by wind through a fuel bed of combustible elements. The numerical model solves a set of time-dependent conservation Equations for both phases (the gas and the vegetation elements) coupled through exchange terms. It accounts for the dynamics, turbulence, soot formation, and radiation. This model has been applied to a prescribed savanna fire. Good qualitative agreement was found between the simulation results and available in situ experimental data on the rate of spread and fuel consumption ratio.

MODELING THERMAL IMPACT OF WILDLAND FIRES ON STRUCTURES IN THE URBAN INTERFACE. PART 1: RADIATIVE AND CONVECTIVE COMPONENTS OF FLAMES REPRESENTATIVE OF VEGETATION FIRES

ABSTRACT A 3-D computational fluid dynamics model is used to estimate the thermal impact on structures exposed to fire in the urban interface. The burning of vegetation is represented by a well-adjusted gas burner diffusion flame. This article examines two situations, depending on which modes of heat transfer from the flame are considered. Model predictions reveal how the presence of the structure modifies the dynamic flow pattern, which in turn may cause the overattachment of the fire plume to the structure and then the enhancement of thermal impact. Numerical results, including flame geometry, are found to be consistent with experimental observations.

DYNAMIC AND RADIATIVE ASPECTS OF FIRE–WATER MIST INTERACTIONS

A two-phase multiclass model is developed to describe the interaction between a compartment fire and a water mist. Turbulent combustion is modeled using the EBU-Ar coupled with the renormalization group k–ϵ turbulence model. A multiphase radiative transfer equation including the contributions of soot particles, combustion products and water droplets is used to model radiation. The model is applied first to a two-dimensional enclosure fire. A parametric study of the influence of water spray flow rate and water droplet diameter on fire mitigation is presented. Gas-phase cooling is found to be the main fire suppression mechanism. The influence of water spray on radiation is studied with special emphasis on the contribution of each radiative phenomenon. Results demonstrated that the attenuation of radiation by the water spray is two-fold: the radiant energy emitted by the flame is reduced and this energy is attenuated by water droplets. The role of droplet scattering in the attenuation of thermal radiation in fire–water mist interactions is clearly demonstrated. For the droplet diameters considered, it is found that there are two distinct regimes: a fire extinction regime and a fire enhanced regime. This finding is consistent with previous experiments. Applied to a three-dimensional fire-sprinkler scenario, the model predictions are in good agreement with experimental data.

A physically based model of the onset of crowning

The objective of this study is to investigate the capability of a physical two-phase model to predict the ignition of crown fuels by a surface fire and then to determine the degree of crowning. The model considers the hydrodynamic aspects of the flow and accounts for the basic physicochemical processes resulting from the thermal degradation of organic matter. Turbulence, soot formation, and its impact on radiation are considered in order to improve the physical insight. Calculations have been performed to investigate the effects of crown base height and aerial fuel moisture content on the onset of crowning. Numerical results are found to be consistent with experimental observations and the widely used Canadian Fire Behavior Prediction System classification by crown fraction burned. This model may be used to extend the domain of application of semiphysical theories, e.g., Van Wagners theory, where fuel and environmental factors are generally determined from empirical observations of previous fires. It provides a means of adjusting these factors in other fire situations without requiring additional experiments.

Sensitivity Analysis of a Fire Field Model in the Case of a Large-Scale Compartment Fire Scenario

The objective of this work is to show how a sensitivity study based on a fractional factorial design can be helpful to quantify the impact of parameter variations on model predictions. These parameters have been carefully chosen due to their high variability in fire modeling and the analysis is conducted by simulating a compartment fire with a CFD model. Through a rigorous approach, it is demonstrated that this fractional design composed of eight simulations gives the same information as a standard full design of 64 runs. Physically, it is found that some turbulence and combustion parameters are significant for most the responses.

MODELING THERMAL IMPACT OF WILDLAND FIRES ON STRUCTURES IN THE URBAN INTERFACE. PART 2: RADIATIVE IMPACT OF A FIRE FRONT

ABSTRACT A fast model of radiative impact on structures exposed to a fire front in the urban interface is presented. The front is viewed as a collection of turbulent diffusion flames whose properties (composition and temperature) are taken from a database previously created from a three-dimensional computational fluid dynamics model. Using the gray soot assumption, two gas radiative property models, the spectral line-based weighted-sum-of-gray-gases and a simpler gray gas model, are compared in terms of accuracy and computational time. Applied to the curved fire front propagation, the thermal response of structures is estimated as well as fire safety zones. It is found that radiation can lead to pilot ignition under the action of firebrands or flame contact.

Heat transfer components at the surface of burning thick PMMA slabs

Mass pyrolysis rate is the key parameter to predict fire behavior. It is generally deduced from the energy balance at the surface of the solid material. However, due to lack of knowledge, existing pyrolysis models use simplifying assumptions neglecting all or part of in-depth losses into the solid material or the net radiation at its surface. In order to improve the accuracy of pyrolysis models, experiments are conducted to quantitatively evaluate the heat transfer components at the surface of burning thick clear poly-methyl-methacrylate (PMMA) slabs at steady state. The contributions of each transfer mode including radiation and convection from the flame, surface re-radiation, and in-depth losses, to total heat flux are determined from two series of experiments. Pure pyrolysis (non-flaming) cone calorimeter experiments are first carried out to evaluate in-depth losses in horizontally-oriented slabs exposed to an incident heat flux below that of ignition. A specific procedure based on video processing is used to track the position of the PMMA regressing surface with time. The second series of experiments consist in burning vertically-oriented slabs from 2.5 cm to 20 cm in height, 10 cm in width and 3 cm in thickness. It is found that only a small part of flame radiation is transmitted through the virgin solid, most in-depth radiation being absorbed by the bubble surface, which in turn strongly emits radiation inward. An excellent agreement is obtained between the local mass loss rate deduced from the energy balance and literature data.

Transport and combustion of Ponderosa Pine firebrands from isolated burning trees

A numerical model is developed to describe the transport and the combustion of firebrands lofted by a fires buoyant plumes. A preliminary study of the thermal degradation and combustion of woody fuel particle is presented. The comparison with lab-scale experiments on cylinder-shaped limbwood samples of Ponderosa Pine (PP) shows a fairly good agreement. A three-dimensional physics-based is used to predict the steady-state flow and thermal fields induced by a crown fire. Trajectories and burning rates of disk-shaped firebrands lofted by the fire plume, and transported downwind are determined for a fire intensity of 20MW/m and various windspeeds from 10 to 20mi/h. Firebrands of different sizes and densities are launched from a specified location at the top of canopy. Results show that the spotting distance depends to the product rhowood0 timestau (rhowood0 : initial wood density, tau : thickness), and varies almost linearly with wind speed while it is independent of the initial particle diameter.