Application of Artificial Neural Network (ANN) to Predict the Curing Rate Equation Parameters of Rubber Foam

Abstract

Rubber foam is one of the materials that is being developed due to its unique properties, such as low density, flexibility, energy absorption ability, etc. Parameters such as compound formulation and vulcanization conditions are affected on these properties. For this reason, kinetic modeling of the vulcanization process is important. In this study, a new method for calculating the parameters which are effect on one of the kinetic models (the Qureshi model) is presented. A model of feed forward back propagation artificial neural network (FF BP ANN) is designed that can predict two model parameters by taking the normalized degree of cure data. The ANN model was trained with the help of this kinetic model data. The capability of the trained ANN was investigated with calculating the error in predicting these parameters. It was found that the parameters were predicted with well accuracy. The experimental data obtained from the rheometer were given as input to the ANN model and the parameters (n and k, the degree and rate constant respectively) related to these data were calculated. This work was also done using the conventional curve fitting method. A comparison between the experimental data reviled that the results obtained from the curve fitting method and the new method presented in this work provides appropriate results in predicting the parameters.