Application of Artificial Neural Network Approach for the Kinetic Parameters Determination of Pyrolysis Aseptic Waste Packages

The thermal degradation behavior for the pyrolysis of aseptic packages was studied using thermogravimetric analyzer under nitrogen atmosphere at the temperature range of 50–900°C with five different heating rates (10, 20, 30, 50, and 100 °C/min). Artificial neural network (ANN) model was also employed in this study to predict the thermal degradation behavior of the aseptic packages. The ANN model employed in this study can predict the result with an accuracy of 99.9%. Model-free methods such as Kissinger-Akahira-Sunose (KAS), Friedman (FR) and Starink methods were employed to determine the kinetic parameters. The KAS model was the most accurate model among the three model-free methods as it has the lowest percentage error. The activation energy, $E_{a}$ for aseptic packages is in the range of 182-218 kJ/mol.