Pyrolysis of Medium-Density Fiberboard: Optimized Search for Kinetics Scheme and Parameters via a Genetic Algorithm Driven by Kissinger’s Method

Abstract

The pyrolysis kinetics of charring materials plays an important role in understanding material combustions especially for constructions materials with complex degradation chemistry. Thermogravimetric analysis (TGA) is frequently used to study the heterogeneous kinetics of solid fuels, however, there is no agreed method to determine the pyrolysis scheme and kinetic parameters for charring polymers with multiple components and competing reaction pathways. This study develops a new technique to estimate the possible numbers of species and sub-reactions in pyrolysis by analysing the second derivatives of TG (DDTG) curves.The pyrolysis of a medium density fibreboard (MDF) in nitrogen is studied in detail, and the DDTG curves are used to locate the temperature of peak mass-loss rate for each sub-reactions. Then, based on the TG data under multiple heating rates, the Kissinger's method is used to quickly find the possible range of values of the kinetic parameters (A and E). These ranges are used to accelerate the optimization of the inverse problem using genetic algorithm (GA) for the kinetic and stoichiometric parameters. The proposed method and the kinetic scheme found are shown to match the experimental data and are able to predict accurately results at different heating rate better than the Kissinger's method. Moreover, the search method (K-K method) is highly efficient, faster than the regular GA search alone. Modelling results show that as the TG data available increases, the interdependence among kinetic parameters becomes weak and the accuracy of first-order model declines. Furthermore, conducting TG experiment under multiple heating rates is found to be crucial in obtaining good kinetic parameters.

A methodology for predicting and comparing the full‐scale fire performance of similar materials based on small‐scale testing

“…The 2 Reaction Model provides a notably better fit with R 2 = 0.96–0.99. In addition to these mechanisms, a decomposition mechanism consisting of 3 parallel reactions that was proposed by Li et al for MDF was considered. The kinetic parameters of individual reactions determined by these researchers were kept unaltered, while the relative amounts of 3 reactants, which according to Li et al represent cellulose, hemicellulose, and adhesive, were varied to fit the MCC data.…”

Experimental investigation on the self-ignition behaviour of coal dust accumulations in oxy-fuel combustion system

“…(5) could not be determined. Nevertheless, the reaction order only changes the shape of DTG (mass-loss rate) curve, not the location of DTG peak [35,36]. So for simplicity, the same second-order reaction (n c ¼ 2) is assumed for all three coals.…”

Pyrolysis of polyurethane foam: optimized search for kinetic properties via simultaneous K–K method, genetic algorithm and elemental analysis

“…It should be noted that the reaction order, n, is mainly affecting the shape of DTG curves instead of the peak temperatures, so called shape factor . The mass fraction at the DTG peak can be determined by rewriting Equation as: …”

“…Because of the range of kinetic properties, selection of the appropriate set of values is needed. The determination of kinetic properties by graphical techniques such as model-free methods can be difficult and its results are questionable for CFD modelling [11].Therefore, since the late 1990s, with the constant improvement of computational capability, several effective property estimation methods and software such as the genetic algorithms (GA) [12,13], PEAKFIT [14] and least-squares method [15,16] have been developed to allow improved kinetic properties estimation. Subsequently, new decomposition models such as the multi-component model [10] and the distributed activation energy model [17] have also been developed.…”

“…The 'uniqueness' of the kinetic properties for different heating rates has been directly challenged recently [18]. Besides, in order to use the enhanced property estimation methods, a set of search regions need to be specified for the kinetic properties, which are usually user defined or based on literature values which are material specific.Recently, a new technique, so called 'K-K' method [11], is developed to fast and accurately predict the material pyrolysis kinetics. The technique is highly efficient, faster than regular optimized searches alone in kinetic properties determination.…”