Elsevier

Combustion and Flame

Volume 157, Issue 4, April 2010, Pages 715-734
Combustion and Flame

An enthalpy-based pyrolysis model for charring and non-charring materials in case of fire

https://doi.org/10.1016/j.combustflame.2009.12.007Get rights and content

Abstract

In a simulation of a developing fire, flame spread must be properly accounted for. The pyrolysis model is important in this respect. To that purpose, we develop a simplified enthalpy-based pyrolysis model that is extendable to multi-dimensional solid-phase treatments. This model is to be coupled to gas phase turbulent combustion simulations. The description of the pyrolysis process is simplified in order to acquire short simulation times. In this paper, first, the basic thermodynamic description of pyrolysis phenomena is revisited for charring and non-charring materials, possibly containing moisture. The heat of pyrolysis is defined and its relation to the formation enthalpies of individual constituents is explained. Solving only one equation for enthalpy on a fixed computational mesh, provides a useful description of the transport of heat and the pyrolysis process inside the solid material. Models for e.g. char oxidation or complex transport of the pyrolysis gases or water vapour inside the solid material can be coupled to the present model. Next, numerical issues and implementation are discussed. We consider basic test cases with imposed external heat flux to a solid material that can be dry or contain moisture. We illustrate that continuous pyrolysis gases mass flow rates are obtained when a piecewise linear representation of the temperature field is adopted on the fixed computational mesh. With constant temperature per computational cell, discontinuities, with sudden drops to zero, are encountered, as reported in the literature. We show that the present model formulation is robust with respect to numerical aspects (cell size and time step) and that the model performs well for variable external heat fluxes. For charring and non-charring materials, we validate the model results by means of numerical reference test cases and experimental data. By means of a numerical test case, we show that the model, when coupled to CFD calculations, is able to simulate flame spread.

Introduction

In the development of a fire, flame spread always plays a very important role. In numerical simulations, this implies coupling of gas phase ‘Computational Fluid Dynamics’ (CFD) simulations, including turbulent combustion and radiation, to pyrolysis simulations in the solid material. In order to make such simulations possible, it is advantageous to keep the pyrolysis model simple.

During the past two decades, several researchers have developed numerical models for pyrolysis of charring materials, with different levels of complexity, such as: Arrhenius-type models [1]; ‘integral’ models [2], [3], [4], [5]; an ‘extended’ integral model [6]; a moving mesh model [7]; a dual mesh model [8]. A review on pyrolysis modelling has recently been provided in [9]. An interesting paper on pyrolysis modelling is reference [10]. For non-charring materials, it is common practice to work with a ‘heat of gasification’ at the pyrolysing surface and to consider a conduction problem in the solid materials (e.g. [2], [11], [12], [13], [14], [15]).

In the present paper, we describe in detail a simplified model, which is applicable to charring and non-charring materials, which can contain moisture. We also explain that the model is extendable for multi-dimensional solid-phase treatments, as required for general flame spread simulations. This is a difference to many existing simplified pyrolysis models, which might look very similar to the present model at first sight, but which are basically limited to one-dimensional configurations (or, at least, implementation for multi-dimensional solid-phase treatments becomes very cumbersome). Also, the present model can be combined with any model for the transport of gases or water vapour inside the solid material. In the same sense, e.g. a model for char oxidation can be added. We consider this beyond the scope of the present paper.

We deliberately avoid inclusion of pyrolysis kinetics. Whereas this limits the field of application to high-temperature pyrolysis and to situations where pyrolysis kinetics is not prevalent, it is relevant for flame spread situations as in a developing fire.

There are two major parts in the model description:

  • the local relation between enthalpy and temperature;

  • the energy equation to be solved in the solid material.

We first describe these parts in detail and define precisely what we mean by the ‘heat of pyrolysis’, in order to avoid any confusion on this term. Afterwards, we put our model and terminology in perspective with respect to existing models in the literature.

Next, numerical issues and implementation, including the solution procedure, are described in detail. An interesting feature of the model is the use of a fixed computational mesh.

Finally, we illustrate the accuracy of the results by means of a series of basic test cases. We discuss the following features:

  • comparison to numerical reference results [5] and experimental data [16] for one-dimensional configurations in charring materials;

  • the importance of the use of a piecewise linear approximation of the temperature field in the solid material for charring materials;

  • sensitivity of the results with respect to the grid size and the time step;

  • comparison to results [2] for PMMA;

  • illustration that the model can deal with moisture in the solid;

  • illustration that the model is applicable to multi-dimensional configurations, as required for general flame spread simulations.

The complete set of results aims at illustrating the robustness and accuracy of the simple model, with applicability to flame spread simulations in a developing fire.

Section snippets

Thermodynamics: introduction

In our model, we focus on the thermodynamic description of the phenomena. Our approach is largely based on [17], one of the first theoretical papers on this topic. Below, we elaborate this to obtain an easy-to-use enthalpy-based pyrolysis model for three-dimensional simulations on a fixed computational mesh. The approach is to consider five constituents: virgin solid material, char, volatiles, liquid water and water vapour. In [17], the endothermic pyrolysis process is assumed to take place at

Implementation and solution procedure

As described in the previous section, the model considers enthalpy as the basic variable, for which a transport equation is solved. We are, however, not interested in enthalpy itself, but rather in temperature distribution and volatile production. The latter is related to the motion of the pyrolysis front, which is assumed infinitely thin in the present model formulation. So, we require a procedure to reconstruct temperature and front position from the basic enthalpy variable. This is done,

Discussion of results

We restrict ourselves to configurations where the externally imposed heat flux is not computed from flame radiation, in order to avoid related uncertainty.

As initial condition, there is only virgin material at temperature T = Tamb = 300 K, which is well below the pyrolysis temperature. Unless stated otherwise, all results are obtained with the piecewise linear temperature field representation.

Summary and conclusions

Starting from a basic thermodynamic description of pyrolysis phenomena, a simplified pyrolysis model was described in detail. The basic model quantity is enthalpy, computed from the specific enthalpies of five constituents (dry virgin material, char, pyrolysis gases, liquid water and water vapour). The concept of heat of pyrolysis and its relation to formation enthalpies of individual constituents was revisited. It was explained how the developed model takes advantage of the use of pyrolysis

Acknowledgments

This research is funded by project G.0130.06 of the Fund of Scientific research – Flanders (Belgium) (FWO-Vlaanderen). The fourth author is Postdoctoral Fellow of the Fund of Scientific research – Flanders (Belgium) (FWO-Vlaanderen).

References (28)

  • C. di Blasi

    Combust. Flame

    (1994)
  • M.J. Spearpoint et al.

    Combust. Flame

    (2000)
  • Z. Yan et al.

    Fire Saf. J.

    (1996)
  • C. di Blasi

    Prog. Energy Combust. Sci.

    (1993)
  • C. Lautenberger et al.

    A model for the oxidative pyrolysis of wood

    Combust. Flame

    (2009)
  • A. Tewarson et al.

    Combust. Flame

    (1992)
  • T. Kashiwagi et al.

    Combust. Flame

    (1987)
  • H.C. Kung

    Combust. Flame

    (1972)
  • B. Fredlund

    Fire Saf. J.

    (1993)
  • E.G. Brehob et al.

    Fire Saf. J.

    (2001)
  • B. Moghtaderi et al.

    Fire Mater.

    (1997)
  • F. Jia et al.

    Fire Mater.

    (1999)
  • W.G. Weng et al.

    Fire Mater.

    (2007)
  • E. Theuns et al.

    Fire Mater.

    (2005)
  • Cited by (30)

    • Pyrolysis and autoignition behaviors of oriented strand board under power-law radiation

      2022, Renewable Energy
      Citation Excerpt :

      Char oxidation, which should be considered in air atmosphere, was not involved in Ira and McKinnon's works, and the utilized constant HFs was a rough approximation in practical fires. Piloted or auto-ignition of solid under constant HF has been studied for decades and different influential aspects were explored, including ambient pressure and oxygen concentration [14], porosity of material [15], generated char layer [16], grain orientation of wood [17], glowing and flaming ignitions [18,19], melting behaviors [20], air flow on surface [21], microgravity [22], in-depth absorption [23–25], optical and radiative properties of semitransparent materials [26]. When exposed to time-dependent HF, solid combustibles may exhibit unique ignition and combustion behaviors.

    • Handbook of Flame Retardants

      2021, Handbook of Flame Retardants
    • Auto-ignition of thermally thick PMMA exposed to linearly decreasing thermal radiation

      2020, Combustion and Flame
      Citation Excerpt :

      Under constant HF, the pioneers did some remarkable workS to reveal the ignition mechanism, such as the classical ignition theory [16] in which critical temperature was used. Based on this original theory, other researchers improved the model when dealing with some influential aspects, such as the surface heat loss [17,18], porosity of materials and the mass transfer of the yielded gas in solid [19,20], grain orientation of wood [21,22], glowing and flaming ignition [22,23], ambient pressure and oxygen concentration [19,20,24–26], air flow velocity on surface [27], generated char layer [28], melting behaviors [29], microgravity [30], in-depth absorption of thermal radiation [1–4], optical and radiative properties of semitransparent materials [31,32]. In order to introduce the critical mass flux into analytical model, Lautenberger [5] and Snegirev [33] used a power law function and Frank-Kamenetskii decomposition [34] respectively to replace the Arrhenius pyrolysis rate and derived explicit expressions for ignition time.

    View all citing articles on Scopus
    View full text