Application of Neural Networks in Vapor-Liquid Equilibrium Estimation of Difluoromethane and 1,1,1,2-Tetrafluoroethane

This paper proposes neural network (NN) based method for calculations of Vapor-liquid equilibrium (VLE) of the binary mixture of Difluoromethane and 1,1,1,2-Tetrafluoroethane. The method is compared with a recently proposed EOS based algorithm. Comparison of these two methods and experimental data shows that NN performs VLE calculation with merely same precision as the EOS method. The main advantage of the proposed NN over the EOS technique is its quicker non-iterative computation. This characteristic makes the NN method more appropriate where real time optimization or model-based control is involved.

کلیدواژه ها:

Neural Networks ، Equation of State ، Vapor-Liquid Equilibrium