A combined heuristic optimization algorithm and Gaussian deconvolution procedure for the kinetics extraction of building wood panel pyrolysis

The thermal degradation behavior of an engineered wood panel, medium density fibreboard (MDF), was experimentally and numerically investigated in the present study. A new strategy for extracting the reaction chemistry kinetics of multiple components and possible overlapping reaction profiles was put forward. The subreactions of each pseudocomponent were separated from the overall thermogravimetric curves, during which a Gaussian deconvolution procedure was employed and the MDF pyrolysis process was identified to involve at least 4 subreactions associated with resin, hemicellulose, cellulose and lignin. After separation, a four component parallel reactions scheme was proposed, in which the kinetic values of each separated decomposition step were further calculated via an isoconversional Advanced Vyazovkin method. The averaged Eα and lnA values are found to be 119.8, 158.5, 179.5, 220.8 kJ/mol and 25.88, 31.05, 33.30, 36.99 s−1 for these pseudo components. The Particle Swarm Optimization algorithm was subsequently utilized to provide the optimization solution for the established four component parallel reactions model under 10, 20 and 30/min simultaneously, in which the initial guesses and the search ranges of the targeted parameters were derived from the Advanced Vyazovkin method estimation results. The optimized Eα and lnA values for resin, hemicellulose, cellulose and lignin are 110.40, 71.06, 137.09, 183.39 kJ/mol and 20.83, 8.19, 22.00, 35.07 s−1, respectively. As well, the proposed strategy and the optimized kinetic parameter values characterizing MDF degradation were validated by inverse modeling simulation of mass loss rate profiles under 40 K/min. Results showed that the simulated thermogravimertric profiles fit well with experimental data under all heating rates, thus proving the practicability and reliability of the proposed numerical strategy.