基于量纲分析的绝热毛细管壅塞流量特性关联  

Dimensional Analysis and Neural Network Correlation of Choked Flow Rate through Adiabatic Capillary Tube

作　　者：张春路[1] 丁国良[1] 

机构地区：[1]上海交通大学制冷与低温工程研究所,上海200030

出　　处：《应用基础与工程科学学报》2002年第1期88-93,共6页Journal of Basic Science and Engineering

基　　金：国家重点基础研究发展规划项目 (G2 0 0 0 0 2 63 0 9)

摘　　要：关联方法可替代复杂的数学模型 ,用以预测绝热毛细管的壅塞流量特性 .先通过量纲分析方法给出了绝热毛细管壅塞流量特性的一般函数关系式 ,然后采用两种模型近似该一般函数关系式 :一种是传统的power law关联式 ;另一种是人工神经网络 .以文献中的R 40 7C毛细管实验数据为对象 ,对比研究了这两种模型的关联效果 .结果表明 :人工神经网络的关联误差及误差的分散度更小 .Instead of the complicated mathematical model, the correlation method could be used to predict the choked flow rate through adiabatic capillary tube. In this work, the dimensional analysis method is used at first to develop a general function. Then, two correlation models are introduced as alternatives of the general function. One is the traditional power law correlation; the other is the forward neural network. The published experimental data of R 407C flowing through adiabatic capillary tubes are correlated with the two kinds of correlation models. It is shown that less correlation error will be got with the neural network method.

关 键 词：毛细管 量纲分析 关联式 人工神经网络 流量特性 关联误差 制冷空调 

分 类 号：TB657[一般工业技术—制冷工程]