物理法再生涤纶鉴别方法初探

初步研究了物理法再生涤纶和原生涤纶的鉴别方法。通过“溶解-沉淀法”提取涤纶中游离齐聚物,用高效液相色谱(HPLC)法检测聚酯齐聚物的分布;在HPLC数据预处理中,依次采用小波变换法、相关性优化规整法和主成分分析方法达到基线校正、信号对齐和数据降维目的;建立反向传播人工神经网络模型,实现物理法再生涤纶的自动识别。该模型不能识别化学法再生涤纶,仅对物理法再生涤纶有效。研究结果表明:采用化学模式识别方法处理涤纶中齐聚物分布可用于鉴别物理法再生涤纶。

Preliminary Research on the Identification of the Physical Recycling Polyester Fibers

This preliminary study aimed to identify and discriminate between the physical recycling polyester fibers and virginal polyester fibers.The oligomers in the polyester fiber samples were extracted by the "dissolution-precipitation" method,and were detected by high performance liquid phase chromatographic(HPLC) method.The HPLC data were successively pretreated for baseline correction,signal alignment and high-dimensional reduction,respectively using wavelet analysis,con-elation optimized warping(COW) and principal components analysis(PCA).The back propagation artificial neural network(BP-ANN) was performed to identify the physical recycling polyester fibers.The BP-ANN failed to predict the chemical recycling polyester fibers,whereas precisely predicted the physical recycling polyester fibers.The results showed that the oligomer distribution assisted by chemical pattern recognition techniques was a potential strategy for the identification of the physical recycling polyester fibers.