利用近红外光谱与族类独立软模式方法（SIMCA）对不同牌号的乙烯醋酸乙烯酯（EVA）共聚物树脂进行快速定性分析。利用漫反射测量方式与傅里叶近红外光谱仪测量样本的近红外光谱，并利用SIMCA方法对训练集样本建立类别定性模型及预测验证集样本的牌号。结果表明，近红外光谱经过导数与MSC预处理后，利用主成分分析方法将光谱数据降维得到光谱的主成分信息，选用前3主成分建立的EVA树脂牌号的定性识别模型，对24个验证集样本的预测准确度为100%。基于近红外光谱与SIMCA定性方法可以用于EVA树脂牌号的快速识别。

Near infrared spectroscopy (NIR) and soft independent modeling of class analogy (SIMCA) are developed in rapid qualitative classification on different grades of EVA resins.76 EVA samples, including Grade EVA2022, EVA2045, EVA3045 are collected from production lines.Of which, 52 samples are used for correction and 24 samples are used for validation.NIR of samples are measured by diffuse reflectance measurement method and Fourier transform spectrometer.SIMCA is used for building qualitative models and predicting the class of validated set.Spectra of all samples are pre-processed by derivation and Multiplicative Scatter Correction (MSC), and are then reduced into low dimensions using Principle Component Analysis (PCA).The first three principle components (PC) are used to establish qualitative distinguishing models for grades of EVA resins.Results show that the corrected rate for class prediction is 100% and the qualitative method is applicable for identifying grade of EVA resins.

谢锦春(1978-),博士研究生,主要从事近红外光谱在聚合物分析表征中的应用研究,xie_jinchun@126.com。