近红外光谱法快速检测宽熔融指数分布的聚丙烯树脂

Rapid detection of polypropylene resins with wide melt flow rate distribution by NIR spectroscopy

熔融指数(MFR)是聚丙烯(PP)的重要加工性质参数,而MFR的传统测量方法无法满足PP聚合过程控制的在线检测需求。针对160个PP粒料样本,采用近红外(NIR)光谱,结合簇状独立软模式法(SIMCA)和偏最小二乘法,建立了判别MFR分布区间和预测MFR的模型。实验结果表明,将PP粒料区分为MFR(10 min)小于1 g、介于1~10 g和大于10 g的样本区间,分别建立NIR光谱预测模型,可提高模型的预测准确性,通过光谱组合预处理,上述3个区间模型的校正标准偏差分别为0.001,0.111,0.691,预测标准偏差分别为0.010,0.766,2.877。使用SIMCA模式识别方法,可准确判定PP样本的MFR数值分布区间,模型的判别准确率为100%。因此可选用相应的NIR光谱校正模型实现宽MFR分布PP粒料的快速检测。

Melt flow rate(MFR)is an important processing property parameter of polypropylene(PP).However,the traditional measurement methods for MFR can not meet the demand for rapid online detection for PP polymerization process control.For 160 PP particle samples,a model for distinguishing the distribution range of MFR and predicting MFR was established by near-infrared(NIR)spectroscopy,combined with soft independent modeling of class analogy(SIMCA)and partial least squares method.The experimental results show that classing the PP pellets into sample intervals with MFR(10 min)less than 1 g,between 1-10 g and more than 10 g to separately establish prediction models by NIR spectroscopy can obviously improve the predictive accuracy of the model.Through spectral combination preprocessing,the corrected standard deviation of the above three interval models is 0.001,0.111,0.691,and the standard error of prediction is 0.010,0.766,2.877,respectively.Moreover,the distribution interval of the MFR values can be accurately discerned by using SIMCA,and the accuracy is 100%.Therefore,for the PP pellets with a wide MFR value distribution,the MFR can be rapid detected by the corresponding NIR spectral calibration model.