摘要
利用神经网络算法推算了超额吉布斯自由能GEm,并将推算结果与实验结果进行了比较,推算的平均相对误差均在1.42%以内。
Correlation and representation of the excess Gibbs energy （Gm^E） with the temperature（T）, pressure （p）, composition（x） and activity coefficient（y） of a system by using neural network have been studied in this paper. Predictions of the excess Gibbs energy based on the trained neural network models have been compared with experiments. The inaccuracy of the predictions was lower than 1.42%.
出处
《新乡学院学报》
2008年第4期27-29,共3页
Journal of Xinxiang University
关键词
超额吉布斯自由能
神经网络算法
自由能参教
excess Gibbs energy
neural network correlation
parameter of free energy
