USE OF NEURAL NETWORKS FOR PREDICTION OF VAPOR-LIQUID-EQUILIBRIUM K-VALUES

This paper presents a new approach for predicting K-values. This approach is based on the use of Neural Networks. The most accurate of existing methods use equations of state. The new approach shows significant improvements in accuracy over equations of state for binary methane and carbon-dioxide mixtures. The Neural Networks method is well adapted to digital computing and results in faster computations than the equation-of-state method. This paper describes Neural Networks and shows how they can be used to predict K-values.

USE OF NEURAL NETWORKS FOR PREDICTION OF VAPOR-LIQUID-EQUILIBRIUM K-VALUES Academic Article