摘要
为了探究化学法再生涤纶和原生涤纶间的差异,应用"醇解-过滤"的方法将聚酯大分子异质链节解聚出来,并经高效液相色谱测试;测试数据经基线校正、信号对齐和数据降维后,转变为两类纤维的特征向量;建立反向传播人工神经网络模型,实现化学法再生涤纶的计算机自动识别。结合所建立的物理法再生涤纶鉴别方法,可以建成一个再生涤纶"两步法"鉴别流程。尽管其存在步骤繁琐的缺点,但势必启发出更为高效的再生涤纶鉴别方案。
In oMer to identify the chemical recycling polyester fibers, the alcoholysis products of polyester heterogeneous chain elements were dissociated from the chemical recycling polyester fibers and virginal polyester fibers through the ＂dissolution- precipitation＂ method, subsequently were measured on high per- formanee liquid chromatography （HPLC） instrument. After baseline correction, signal alignment and high- dimensional reduction, the HPLC data were transformed into characteristic vectors of the two fiber species. By building a preliminaJ.＇y model of back propagation artificial neural network （BP-ANN）, the chemical recy- cling polyester fibers could be identified automatically on computer. Combined with the identification meth- od of the physical recycling polyester fibers established by the author, a ＂two step＂ authentication process of regenerated polyester fibers was performed. Although there were defects in convenience, this preliminarily study surely wouhl inspire a more efficient solution to the authentication of regenerated polyester fibers.
出处
《合成纤维》
CAS
2016年第8期41-45,共5页
Synthetic Fiber in China
关键词
化学法再生涤纶
鉴别
异质链节
化学模式识别
反向传播人工神经网络
chemical recycling polyester fiber, identification, heterogeneous chain element, chemical pat- tern recognition, back propagation artificial neural network
作者简介
作者简介：付昌飞，男，博士，主要从事合成纤维改性及表征工作
