基于小波神经网络的精梳毛纱条干不匀预测(英文)  

Application of a wavelet neural network on the prediction of worsted yarn evenness

作　　者：王侃枫[1] 姜景连[2] 马兵 

机构地区：[1]东华大学纺织学院,上海200051 [2]武夷学院数学与计算机系,福建武夷山354300 [3]美乐家(中国)日用品有限公司,上海200072

出　　处：《纺织高校基础科学学报》2011年第1期136-140,共5页Basic Sciences Journal of Textile Universities

摘　　要：运用一种小波神经网络预测精梳毛纺纱线的条干不匀.通过分析纺纱过程,选取合适的影响参数,建立组合神经网络测试纱线的CV%,细节,粗节和毛粒.预测值和实际值之间的相关系数的平方分别为0.9854,0.9758,0.9312和0.8474.表明小波神经网络是预测纱线条干不匀的一种有效方法.A wavelet neural network（WNN） was proposed and applied to the prediction of the worsted yarn evenness.A compound network was constituted to predict yarn CV%,the numbers of thin places,thick places and neps by analyzing the spinning theory and choosing the right input parameters.Excellent predictions results with a square of correlation coefficient of 0.985 4,0.975 8,0.931 2 and 0.847 4 respectively for the four indexes suggested that the WNN was a prospective mothod in prediction of the yarn evenness.

关 键 词：小波神经网络 纱线条干不匀 纱线CV% 细节 粗节 毛粒 

分 类 号：TS101.922[轻工技术与工程—纺织工程]