Multi-scale modeling of the thermal decomposition of fire retardant plywood

• TGA analysis was used to propose a mechanism of the thermal decomposition of plywood.
• Cone calorimeter was used to study the thermal decomposition of plywood.
• Thermal properties calculated by using GA method.
• GPYRO code was used to modeling the cone calorimeter experiments.
• Mass loss, mass loss rate and temperature profile have been determined for different levels of heat flux.

Due to the complexity and costs of full scale fire-test experiments, numerical simulations provide a useful alternative when investigating the fire behavior of new materials. The mass loss rate of the solid is one of the most important parameters in assessing fire behavior as it is directly linked with the pyrolysis gas flow rate and represents the initial factor of the combustion process. In this paper, fire retardant plywood is investigated with a focus on the solid mass loss rate modeling. A multi-scale approach is followed in order to establish the kinetic mechanism of thermal degradation. A combination of small scale and large scale tests were completed to fully develop and validate the proposed kinetic mechanism. For small scale testing, experiments are conducted by using thermo-gravimetric analysis coupled to gas analysis with FTIR technique under nitrogen and air atmospheres. These experiments were completed at several heating rates. Thermo-gravimetric results are used to propose a kinetic mechanism for the thermal decomposition of the solid and the kinetic parameters are calculated by using the genetic algorithms method. For larger-scale testing, experiments were carried out in a cone calorimeter coupled to a FTIR gas analyzer. The experiments were completed in air atmosphere in order to validate the kinetic mechanism developed from small-scale testing. The kinetic model developed is implemented into the general Gpyro model which takes into account both thermal and mass transfer phenomena inside the solid. The results showed good agreement between the model calculations and the experimental data.