利用神经网络在线预测香菇干燥含水率  

Moisture Content Prediction of Xianggu Mushroom Drying Based on ANN

作　　者：付立思[1] 孙晓杰[1] 杨勇[1] 朴在林[1] 

机构地区：[1]沈阳农业大学信息与电气工程学院,沈阳110161

出　　处：《沈阳农业大学学报》2005年第6期663-666,共4页Journal of Shenyang Agricultural University

基　　金：辽宁省科技攻关资助项目(2002206001)

摘　　要：目前的检测方法与检测仪器,难以实现香菇等农副产品干燥过程中含水率的在线检测。影响干燥过程中香菇在线含水率的因素包括各干燥阶段的时间、温度、通风量以及初含水率和所处的干燥阶段。探讨了基于神经网络的香菇干燥含水率在线检测的方法。确定了以干燥各阶段时间、温度、通风量和所处干燥阶段为系统输入,有偏差的两层BP神经网络。利用M at-lab对系统进行建模,干燥第一阶段进行的仿真结果表明,在两个训练单位时间里,达到了均方误差要求。It is difficult for predicting the moisture content of agricuhural byproducts during drying period on-line with current measuring methods or instruments. The factors that affect the moisture content of drying Xianggu mushroom include time of the different drying phases, temperature, ventilation flux, original moisture content and the drying phase. The detection method of Xianggu mushroom drying based on ANN was discussed. The system with drying time, temperature, ventilation flux and drying phase as the inputs comprises two layers of BP-ANN. The simulating results of the first phase showed that the mean square error reached the required level in two training time units.

关 键 词：香菇干燥 含水在线检测 神经网络 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置]