摘要
采用平衡法,测定三聚氰酸在水中的溶解度,其数据用Apelblat溶解度方程关联的结果为:lnx=-474.8025+18932.9494T+70.7514lnT,计算值与实验值、文献值间的平均相对误差为1.28%.以温度为输入矢量、溶解度为输出矢量,建立RBF神经网络,结果表明,RBF神经网络进行函数逼近可实现网络的快速收敛,训练集平均误差为0.81%,测试集平均误差为1.64%.因此,所建立的RBF人工神经网络模型对三聚氰酸-水体系在283.15~363.15K间是适用的.
Solubility of cyanuric acid in water systematically is determined. By Apelblat's solubility equation correlating, the equation lnx=-474.802 5+18 932.949 4T+70.751 41lnT is obtained. The mean relative deviation is 1.28% between the calculated values and the experimental, literature values. Taking temperature as input vector and solubility as output vector to set up radial basis function artificial neural networks, the results indicate that RBF can fast realize convergence approaching a function. The mean relative deviation of train aggregate is 0.81%, and test aggregate is 1.64%. The RBF neural network is suitable for the system of cyanuric-water from 283.15 K to 363.15 K.
出处
《郑州大学学报(工学版)》
CAS
2003年第2期19-21,32,共4页
Journal of Zhengzhou University(Engineering Science)
基金
1999年度河南省杰出青年科学基金资助项目(9909)