论文部分内容阅读
面向对二甲苯(PX)成品的在线分析问题,为克服色谱分析法的不足,本文提出了一种基于拉曼光谱的PX成品杂质含量分析方法。为了克服激光中心波长与功率波动带来的不良影响,提出了激光中心波长自动校正与PX特征峰归一化方法;同时,结合相关分析获取了主要杂质的特征波段,并采用偏最小二乘算法进行定量分析。针对11个未知样本,PLS模型对甲苯(MB)、乙苯(EB)、间二甲苯(MX)、邻二甲苯(OX)的标准预测误差分别为0.03%、0.11%、0.02%、0.03%。实验结果表明,拉曼光谱法能够快速、准确且无损地测定PX成品中主要杂质的含量,为在线拉曼分析系统的开发奠定了方法基础。
In order to solve the problem of on-line analysis of p-xylene (PX) finished products, this paper presents a method for the analysis of impurities in PX finished products based on Raman spectroscopy. In order to overcome the adverse effect of laser center wavelength and power fluctuation, a method of laser center wavelength automatic correction and PX peak eigenvalue normalization was proposed. At the same time, the characteristic bands of main impurities were obtained by correlation analysis, and partial least squares Quantitative analysis. For the 11 unknown samples, the standard prediction errors of PLS for toluene, ethylbenzene, m-xylene and o-xylene were 0.03%, 0.11%, 0.02%, 0.03% . The experimental results show that Raman spectroscopy can determine the content of major impurities in the finished product of PX fast, accurately and non-destructively, which lays a foundation for the development of online Raman analysis system.