PREDICTION OF VAPOR-LIQUID EQUILIBRIUM FOR QUATERNARY SYSTEM BASED ON THE CONSTITUENT TERNARY SYSTEMS DATA BY USING BP ARTIFICIAL NEURAL NETWORK

摘要

Choosing experimental vapor-liquid equilibrium (VLE) data of four constituent ternary systems as training samples of BP artificial neural network, liquid mole fractions as input, and vapor mole fractions and equilibrium temperatures as output respectively, BP neural network prediction model A of VLE for methyl acetate--methanol--water--acetic acid system at 101.3kPa was built. 19 VLE data of the quaternary system were predicted and the results were satisfactory. Similar to model A,BP neural network prediction model B of VLE for acetone--methanol--2-propanol--1-propanol system at 101.3kPa was built. 10 VLE data of the quaternary system were predicted and the prediction results were in very good agreement with the experimental data. The results show that BP neural network is an effective method in VLE prediction of quaternary system.