中药提取软测量系统的设计与实现  

Design and implementation of a soft sensor system for Chinese medicine extraction

作　　者：齐岩磊[1] 陈娟[1] 杨洋[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029

出　　处：《北京化工大学学报（自然科学版）》2011年第5期130-135,共6页Journal of Beijing University of Chemical Technology(Natural Science Edition)

基　　金：国家自然科学基金(20676013)

摘　　要：针对超声波提取中草药活性成分含量检测困难的问题,通过采用可测量的辅助变量建立基于神经网络BP算法的软测量模型来估计不可测量变量的值,并利用单片机进行辅助变量的数据采集和软测量模型计算,实现了用软仪表对超声波中药提取率的在线测量。与目前采用紫外分光仪离线取样进行检测相比,该模型克服了离线测量工作量大、无法直接得到提取率的测量值等缺点,具有测量精度较高(测量相对误差控制在5%以内)、响应速度快和实时性好等特点。Since the detection of the active components in Chinese medicine is difficult, a soft sensor model based on a neural network BP algorithm has been established in order to estimate the values of variables that cannot be obtained through measurable auxiliary variables. Using a microcontroller unit for data collection and soft sensor model calculations, the on-line detection of the rate of extraction of medicinal components by ultrasonieation has been implemented with virtual instruments. The soft sensor system overcomes the shortcomings associated with the currently used method of off-line sample detection by UV spectrometry, namely the large workload and inability to directly measure the extraction rate. The new system has the following advantageous features： high accuracy （relative errors within 5% ） , good generality, fast response and good real-time performance.

关 键 词：单片机 软测量仪表 神经网络 提取率检测 

分 类 号：TP212.2[自动化与计算机技术—检测技术与自动化装置]