A New Enhanced Strategy Based on MLP Neural Network For CalculatingCoefficient in Dense-diluted Liquids

A new optimum, mathematically developed correlation has been proposed that strictly applicable to estimate diffusion coefficient in diluted solutions. Comparison has been done between the earlier most used correlation, Wilke, and new developed correlation. Amount of deviation of results in compare to experimental data has been calculated. Variety ranges of data points (upper than 280 points in various conditions of pressure and temperature) which are cited in previous experimental works have been used as a general major empirical data source for developing process. Artificial Neural Network has the ability to identify nonlinear dynamic systems fromtraining data. This paper describes the use of ANN for predicting the diffusion coefficient in dense-diluted liquids. In our approach, the input descriptors obtained from experimental data. The results extracted from MLP neural network showing that our method is better than other formulas and methods

A New Enhanced Strategy Based on MLP Neural Network For CalculatingCoefficient in Dense-diluted Liquids Keywords:

molecular diffusion , liquid , empirical data , Wilke , MLP , Neural Network