基于BP神经网络的机采原棉品质指标预测模型  

Machine Pick up Cotton Quality Index Prediction Model Based on BP Neural Network

作　　者：李勇[1] 陈晓川[1] 汪军[2] 张洪洲[1] 王伟[1] 

机构地区：[1]塔里木大学,新疆阿拉尔843300 [2]东华大学,上海201620

出　　处：《棉纺织技术》2015年第1期17-20,56,共5页Cotton Textile Technology

基　　金：新疆生产建设兵团支疆计划项目(2012AB008);国家质检总局项目(201310107);塔里木大学校长基金自然科学项目(TDZKSS201322)

摘　　要：为了对脱籽后的机采原棉品质指标进行预测并实现优化控制,设计了用于机采原棉品质指标预测的BP神经网络模型。以南疆地区机采棉为研究对象,以影响原棉品质的主控因素籽棉回潮率和轧花速度为BP神经网络模型的基本特征量,建立了机采原棉品质指标的BP神经网络预测模型。结果表明:该BP神经网络模型能较好表达机采原棉各品质指标与主控因素之间的非线性关系,预测结果与实测值之间误差小,测试样本的网络输出值与网络目标值的相关系数均接近1,模型预测效果较佳。认为:该BP神经网络模型可作为机采原棉品质预测与调控的新方法,也可应用于机采籽棉轧花在线原棉品质监控。To predict and control quality indexes of machine pick up raw cotton after removal seed,machine pick up cotton quality index prediction model based on BP neural network was designed. Taking Xinjiang south zone machine pick up cotton as research object, seed cotton moisture regain ＆ gin speed that effected raw cotton quality were used as BP neural network model basic characteristic index, BP neural network prediction model based on machine pick up cotton quality index was established. The result shows that nonlinear relationship he tween machine pick up cotton each quality indexes and seed cotton moisture ＆ gin speed can be displayed better by the model. The deviation between prediction result and practical test value is less. The correlation coefficient of test samples network output value and network object values is close to 1,the prediction effect is better. It is considered that the model can be used as new method of prediction and control machine pick up cotton,also can be used for machine pick up seed cotton online detection.

关 键 词：BP神经网络 机采棉 预测模型 上半部平均长度 整齐度指数 短纤指数 

分 类 号：S225.911[农业科学—农业机械化工程] TP183[农业科学—农业工程]