Comparing artificial neural network with conventional kinetic model for investigation of thermal decomposition in nanocomposites

Polyimide-Silica Hybrid nanocomposite samples were prepared by sol-gel technique. Specimens from the hybrid nanocomposite were submitted to thermogravimetric analysis and thermal degradation kinetics of hybrid nanocomposite was investigated by thermogravimetric analysis. The kinetic parameters were obtained via the chemometric data processing of mass loss curves. The non-linear fitting method based on Particle Swarm Optimization (PSO) algorithm was used to fit the mass loss curves at three heating rates and to adjust the non-linear curves. PSO is a population based stochastic optimization technique inspired by social behavior of bird flocking or fish schooling.