红外光谱测定丙烯酸聚合反应的定量研究

Quantitative study on the polymerization of acrylic acid by infrared spectroscopy

为在线定量测定丙烯酸聚合反应进程,本文采用红外光谱分析技术,以偏最小二乘法(PLS)、支持向量机回归(SVR)、人工神经网络(ANN)等方法对丙烯酸和聚丙烯酸建立定量分析模型。提出基于上述定量模型不同组合的集成模型方法,对五组不同加料量的聚丙烯酸聚合反应过程进行定量分析,验证红外光谱定量应用于聚合反应的预测准确性。结果表明红外定量分析方法在聚合反应进程分析中拥有良好的预测性能,不同定量方法中SVR-POLY对丙烯酸的平均预测误差最小(0.0022±0.0021),PLS、SVR-POLY、SVR-RBF、ANN组合的集成模型对聚丙烯酸的平均预测误差最小(0.0037±0.0046)。

In order to quantitatively determine the polymerization process of acrylic acid online,a quantitative analysis model of acrylic acid and polyacrylic acid was established by using infrared spectroscopy,Partial Least Squares(PLS),Support Vector Regression(SVR),Artificial Neural Network(ANN)and other methods.An ensemble model method based on the above quantitative model is proposed to quantitatively analyze the polymerization process of five groups of polyacrylic acid with different feed amounts,and verify the prediction accuracy of quantitative application of infrared spectroscopy to polymerization.The results show that the infrared quantitative analysis method has good prediction performance in the analysis of polymerization process.Among the different quantitative methods,SVR-POLY has the least average prediction error for acrylic acid(0.0022±0.0021).Ensemble Model of PLS、SVR-POLY、SVR-RBF and ANN has the least average prediction error for polyacrylic acid(0.0037±0.0046).