BP神经网络在中药水提液膜过滤中的应用  

Application and Implementation of BP Neural Net in Chinese Herbal Medicine'S Membrane Filtration

作　　者：李玲娟[1] 李刚[1] 

机构地区：[1]南京中医药大学信息技术学院,江苏南京210046

出　　处：《计算机仿真》2009年第6期195-199,共5页Computer Simulation

基　　金：国家自然科学基金项目:(30572374);江苏省高校自然科学研究指导性计划项目(06KJD520144)

摘　　要：为防治中药水提液在无机陶瓷膜过滤中的膜污染问题,采用MATLAB编程实现了BP神经网络算法,并以反映中药水提液膜过程中的特征实验数据为训练样本建立网络模型。应用模型可预测反映水提液过滤时膜污染程度的指标参数——膜通量和膜污染度,从而为膜污染防治提供理论依据。网络的训练速度及预测结果表明,算法收敛速度较快,预测精度达到要求,为中药水提液膜过滤复杂系统中的膜污染程度预测提供了一种新思路和新方法。The present paper aims at solving the probiem of membrane fouling control in the fihration. In this study, the implementation of BP neural net algorithm based on MATLAB programming sets up a network in membrane extracting of Chinese herbal medicine. Taking physical and chemical parameters of water extract as samples, the well -practiced network model was utilized to forecast membrane flux and membrane fouling degree. It showed that with higher convergent speed and reliable forecast accuracy, this algorithm would be a new idea and method to forecast membrane fouling degree in complicated filtration process.

关 键 词：反向传播网络 膜通量 膜污染度 预测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]