测定表面张力的滴体积法校正系数的外推模型  

Extrapolating model of correct coefficient in the dropvolume method for determination of surface or interfacial tension

作　　者：冯玉海[1] 沈本贤[1] 高晋生[1] 

机构地区：[1]华东理工大学能源化工系,上海200237

出　　处：《日用化学工业》2002年第4期4-6,共3页China Surfactant Detergent & Cosmetics

摘　　要：应用神经网络的方法 ,建立滴体积法测定表面张力和界面张力计算公式中的校正系数F与V/R3 的外推模型。用部分已知的F -V/R3 数据 (0 5 44 0 <V/R3 <3 3 5 7)进行训练 ,考察所建立的神经网络模型对未参加训练的V/R3 的F值预测能力。结果表明 ,该方法具有较好的外推性 ,其预测值与文献值的平均相对误差为 0 0 2 4% ,然后用全部的F -V/R3 数据 (0 5 44 0 <V/R3 <3 7 0 4)进行训练 ,建立滴体积法校正系数的神经网络模型 ,并利用该模型对V/R3 >3 7 0 4的F值进行了外推预测 。A extrapolating model of correct coefficient(F)in the dropvolume method for determination of surface and interfacial tension was developed by using neural network.The prediction results of the model got by training part data of F-V/R 3 showed that the method has a good prediction ability and the average relative error of the predicted and the measured is 0 024 % .Therefore,the extrapolating model of correct coefficient(F)was established by training all of the data of F-V/R 3( 0 544 0 <V/R 3<37 04)and the value of F for V/R 3>37 04 were predicted by this model.Then the application range of dropvolume method was extended.

关 键 词：测定 表面张力 滴定积法 校正系数 外推模型 神经网络模型 

分 类 号：O552.421[理学—热学与物质分子运动论]