以二元混合物体系的实验数据为基础，通过神经网络表达混合物组成与过量摩尔体积的定量关系，进而推测三元混合物体系的过量摩尔体积．本文研究了16个极性和非极性混合物体系，对其过量摩尔体积的预测结果优于Rastogi方程的计算结果，表明神经网络可以作为一种推测三元混合物体系过量摩尔体积的工具。 Based on the experimental data of the binary mixture system, the quantitative relationship between the composition of the mixture and the excess molar volume is expressed by the neural network, and then the excess molar volume of the ternary mixture system is estimated. In this paper, 16 polar and non-polar mixture systems are studied, and the prediction of the excess molar volume is superior to that of the Rastogi equation. It shows that the neural network can be used as a tool to estimate the excess molar volume of the ternary mixture.