Prediction of Cell Density of Polystyrene/NanosilicaFoams by Artificial Neural Network

In this work, artificial neural network (ANNs) method is used to make a prediction of cell density of Polystyrene (PS)/nanosilicananocomposite foams. To predict cell density of foam with through this method, we use three variables are used in different levels including interface tension between polystyrene and carbon dioxide (72, 39.5, 29 mJ/m2) size of nanosilica (12, 20, 40 nm) and amount of nanosilica (0.5, 1, 2 wt. %). Results obtained by ANNs method is in very good agreement with the reported experimental data with minimum errors. ANN results confirm that the cell density of foam is increased from 1.99×108 cell/cm3 to 9.52×108 cell/cm3 by decreasing in nanosilica size and increasing in the amount of nanosilica. ANN model R2 for training data and testing data was 1.00 and 0.931, respectively