preparation of expandable polystyrene by multi-stage initiator dosing/styrene-butadiene-styrene blends with application of artificial neural networks

Expandable polystyrene (eps) is one of the most used polymers. preparation of this polymer by the conventional method has some problems which cause the synthesis process to be difficult and also decrease the quality of the prepared eps. in this study, styrene-butadiene-styrene (sbs)  has been added to improve some properties of the prepared polymer and the multi-stage initiator dosing (mid) method has been used to reduce the time of the polymerization which causes the polymer’s production capacity to increase. sbs has been added to eps in shares of 2%wt, 4%wt, and 6%wt. the polydispersity index (pdi) test and the amount of tension in the yield point of the polymer have been checked. the amount of absorbed pentane on the polymer studied. the amount of residual monomer on the polymer has been investigated. all of the studies happened under different conditions like different percentages of initiator, different numbers of dosings, and different time periods of the first stage of the polymerization. experimental data have been simulated by multi-layer perceptron (mlp) and radial basis function (rbf) methods of artificial neural networks (ann). the performance of the simulation for the rbf method was better in comparison to the mlp method due to having a strong scientific foundation and also the ability to filter noises. the experimental data show that a higher amount of sbs causes improvement in properties like elongation at break, better pentane absorption, and pdi amount has improved, which shows the better distribution of molecular weight and a decrease in residual monomer in products.