摘要
使用神经网络模型 ,对纯物质的饱和液体传递性质粘度、导热系数及表面张力与温度的函数关系进行预测。对常见的 35 0多种有机物的预测结果表明 ,在熔点到临界点的温度范围内 ,粘度、导热系数及表面张力的平均预测误差分别为 0 12 %、 0 0 9%及 0 0 3%。
Neural Network model is used to predict transmitting property of the puresaturated liquid such as viscosity, coefficient of heat conductivity and surface tension. Predicting results based on 350 organic compounds showed that the estimated averageerror of viscosity, coefficient of heat conductivity and surface tension are respectively 0 12%, 0 09% and 0 03%between the melting point and critical point.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2001年第5期577-580,共4页
Computers and Applied Chemistry