Toward prediction of surface tension of branched n-alkanes using ANN technique

This study features the application of artificial neural network for prediction of surface tension of branched alkanes. Surface and interfacial tensions of alkanes show a specific interaction between adjacent molecules of the higher n-alkanes which results in an anisotropic dispersion force component of the surface energy. The surface tension of branched alkanes was studied for temperatures between 283.15 and 448.15 K. Two intelligent models named multilayer perceptron model (MLP) and radial basis function (RBF) model were developed and the accuracy of two models was examined by different graphical and statistical methods. Results showed that the both models are accurate and effective in prediction of surface tension of branched alkanes. However, the results were compared with experimental data and it was found that the estimated surface tension by multi-layer perceptron neural network is more accurate than radial basis function network.