论文部分内容阅读
通过分析若干聚乙烯样品的拉曼光谱,对聚乙烯密度的检测方法进行了研究.分别采用拉曼光谱内标法(RISM)和偏最小二乘法(PLS)建立聚乙烯密度预测模型,其中RISM方法建模所得检测结果与真实值的相关系数为0.558,平均相对偏差(ARD)为1.1%,标准偏差(SD)为11.7,而PLS方法建模所得检测结果的R、ARD和交叉验证误差均方根(RMSECV)分别为0.985,0.2%和2.2.研究表明,通过拉曼光谱能够实现聚乙烯密度的快速检测,同时,基于PLS的建模分析方法检测结果更优.
By analyzing the Raman spectra of several polyethylene samples, the density of polyethylene was examined, and the polyethylene density prediction model was established by Raman spectroscopy (RISM) and partial least squares (PLS) respectively. Among them, RISM The correlation coefficient between the test results and the real value was 0.558, the average relative deviation (ARD) was 1.1% and the standard deviation (SD) was 11.7. The R, ARD and cross-validation errors of PLS method were all Root mean square (RMSECV) were 0.985, 0.2% and 2.2, respectively. Studies have shown that rapid detection of polyethylene density can be achieved by Raman spectroscopy, and the PLS-based modeling analysis method has better detection results.