织物柔软度等级评价的BP网络方法  

Fabric Softness Evaluation with BP Neural Network

作　　者：欧建文[1] 李淑红[1] 梁小平[1] 樊小伟[1] 

机构地区：[1]天津工业大学,天津300160

出　　处：《山东纺织科技》2009年第6期39-42,共4页Shandong Textile Science & Technology

摘　　要：文章基于织物柔软度等级评价体系,建立一个织物力学指标与柔软度关系的BP神经网络模型,对织物柔软度进行评价。借助KES-F风格仪测得的12组数据对BP神经网络进行训练,训练好的模型对织物进行检验。结果表明:网络迅速完成训练,误差平方和低于10-3;对检验用的织物进行等级评价,其输出等级与综合评价等级保持一致。此方法客观、准确、简捷。Based on the evaluation system of fabric softness, a BP neural network model of relationships between mechanical index and flexibility of fabrics was built to evaluate fabric softness. The BP neural network was practised with 12 sets of data measured by KES-F-style instrument, and fab- rics was tested with the trained model. The results show that the net complete the training rapidly and the output error is less than 10 ^-3. If it is used to evaluate fabrics rating, the output should consistent with the level of comprehensive evaluation. So this method is objective, accurate and concise.

关 键 词：织物 柔软度 BP神经网络 评价方法 

分 类 号：TS101.9[轻工技术与工程—纺织工程]