节点文献
人工神经网络基团贡献法估算纯有机物的临界参数
Prediction of Critical Properties for Organic Compound by Group-Contribution Artificial Neural Network Method
【摘要】 提出了估算纯有机物临界参数的人工神经网络基团贡献法 ,网络的输入参数为基团和常压沸点 ,临界温度 (Tc)、临界压力 (Pc)与临界比容 (Vc)同时作为网络的输出 ,对 16 5种碳氢氧化合物分别进行了预测 ,Tc、Pc 和Vc 的平均相对误差为 :1 5 6 %、 3 49%和 3 2 1%。与通用的基团贡献法MXXC法进行比较 ,表明人工神经网络基团贡献法更具有优越性
【Abstract】 Based on group contribution,the ANN method was used to predict the critical properties of organic compound. The inputs are groups and T b ,the outputs are T c,P\-c and V c . The average percent deviations of properties of T c,P c and V c are only 1 56%, 3 49% and 3 21% in 165 hydrocarbons. The deviation of Pc, and Vc are lower than those estimated by popular MXXC method chose for comparison. ANN method shows many advantages in prediction of critical properties.
【关键词】 人工神经网络;
临界参数;
估算;
基团贡献;
【Key words】 neural network; critical properties; prediction; group contribution;
【Key words】 neural network; critical properties; prediction; group contribution;
- 【文献出处】 计算机与应用化学 ,Computers and Applied Chemistry , 编辑部邮箱 ,2001年04期
- 【分类号】O621
- 【被引频次】15
- 【下载频次】171