神经元网络软测量模型在中药浓缩工段的应用  

Application of ANN Soft-sensing Model in Traditional Chinese Drug Inspissation Process

作　　者：韦文祥[1] 沈洪远[1] 罗安[2] 

机构地区：[1]湖南科技大学信息与电气工程学院,湖南湘潭411201 [2]湖南大学电气与信息工程学院,长沙410082

出　　处：《化工自动化及仪表》2007年第5期57-61,共5页Control and Instruments in Chemical Industry

基　　金：国家计委中药生产自动化示范工程基金资助项目(计高技[2002]2141)

摘　　要：运用人工神经网络技术建立了在中药浓缩工段对药液浓度进行测量的BP网络软测量模型,利用M atlab 6.5神经网络工具箱对该模型进行了初始化、训练和测试,样本数据来自某中药自动化智能控制系统的浓缩工段,训练采用Levenberg-M arquardt训练方式,结果表明该模型的输出与样本值的一致性较好。　A kind of BP soft-sensing model based on ANN was established and was used to measure the concentration of traditional Chinese medicine in inspissation process.The model was initialized,trained and tested by using NN tools box in Matlab6.5 and the sample data were coming from an automatic intelligent control system of tradition Chinese drug.The training process adopted Levenberg-Marquardt and the results show that output data are consistent with sample data.

关 键 词：中药浓缩 浓度软测量 BF网络 ANN 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]