基于BP神经网络的玉米膳食纤维挤压改性预测模型  

Prediction Model of Corn Quality Extruded by Twin-screw Extruder based on Back Propagation Neural Network

作　　者：邓力[1,2] 何腊平[1,2] 姚翔[1,2] 绕元安 

机构地区：[1]贵州大学生命科学学院,贵州贵阳550025 [2]贵州省农畜产品贮藏与加工重点试验室,贵州贵阳550025 [3]贵阳开元生物技术有限公司,贵州贵阳550025

出　　处：《农产品加工（下）》2014年第3期5-8,共4页Farm Products Processing

基　　金：贵阳市科学计划项目生物重大专项([2010]筑农合同字第8-1号);贵州省科技支撑计划项目(黔科合NY字[2011]3099号);贵州农畜产品贮藏与加工重点试验室建设(黔科合计Z字[2012]4001)

摘　　要：以玉米为原料进行挤压试验,开发基于双螺杆挤压的玉米膳食纤维改性的BP神经网络预测模型。此网络模型以螺杆转速、喂料速度、含水量和机筒温度为输入单元,以糊化度、吸水性(WAI)和水溶性(WSI)为输出单元,拥有一个8单元的隐含层。网络输出和目标输出之间的相关系数为0.984 46,预测误差小于10%,具有较好网络性能,能够实现对玉米糊化度、吸水性和水溶性等挤压性能的预测。进一步开发人机交互图形化用户界面设计(GUI),方便预测模型的应用。Extrusion experiments are carried out using corn as the material to develop a back propagation （BP） neural network prediction model based on corn quality extruded by twin-screw extruder. This network model is to use the screw speed, feed rate, water content and temperature of the barrel as an input unit, the degree of gelatinization, water-absorbent index （WAI）, and water-soluble index （WSI） as an output unit. It has an 8-unit of the hidden layer. The linear regression slope is 0.984 46 between the network output and the target output, and the prediction error is less than 10%. It shows that the network performs satisfactorily, and can be used to predict the degree of gelatinization, WAI and WSI of corn based on the network model. Further, graphical user interface （GUI） is developed to facilitate its application.

关 键 词：玉米 挤压 BP神经网络 预测模型 图形化用户界面设计 

分 类 号：TS213[轻工技术与工程—粮食、油脂及植物蛋白工程]