基于相空间重构的纱条不匀预测  

Phase space reconstruction in sliver evenness series

作　　者：姚杰[1] 叶国铭[1] 侯湘洪[1] 

机构地区：[1]东华大学机械工程学院,上海200051

出　　处：《纺织学报》2005年第4期57-58,61,共3页Journal of Textile Research

摘　　要：根据纱条不匀存在着的非线性、时变性和不确定作用关系,应用局域线性预测法和全局神经网络预测法进行了预测分析。结果表明,该二种方法能够对纱条不匀进行有效的短期预测,同时在该条件下,其中局域线性预测法效果较好,该结果证实了混沌非线性方法在纱条不匀分析中的有效性。According to the nonlinear characteristics of sliver evenness, a local linear forecasting method and a global neural networks forecasting method were given to predict a segment of sliver evenness. The results show that the two methods are good at forecasting the series in short-term, and the local linear forecasting is better than the other in this condition. At the time, it also approves the validity of the application of chaos nonlinear in the analysis to sliver evenness.

关 键 词：纱条不匀 相空间重构 局域线性预测 全局神经网络预测 

分 类 号：TS101.8[轻工技术与工程—纺织工程]