摘要
本文基于正交投影的正交校正算法(OSC),建立了聚氯乙稀近红外光谱与其K值之间关系的数学模型,提出了用近红外漫反射光谱技术快速检测聚氯乙稀K值的新方法。研究结果表明,原始光谱经正交信号校正处理后,聚氯乙稀K值化学测定值与近红外预测值的相关系数和标准差为R=0.9825、RMSEP=0.5933,预测结果的相对误差在0.5%以内,且正交信号校正算法能取得比单纯使用偏最小二乘回归更好的预测效果,模型更为简单。为用近红外光谱快速测定聚氯乙稀的K值提供了一条新途径。
A new method for fast determining the K value, of poly （vinyl chloride） by means of near infrared （NIR） spectroscopy was developed. Orthogonal signal correction method was introduced to model the quantitative correlation between the NIR spectroscopy and the K value of poly （vinyl chloride）. The results show that the correlative coefficients and the standard error of prediction （RMSEP） between the chemical analysis value and the NIR predicated value of the K value of the sample were 0.9825 and 0.5933 after the orthogonal signal correction. The relative error of the predicted results was less than 0.5%. The experiment results show that better performance and simpler partial least square model can be achieved with the orthogonal signal correction algorithm than Partial Least Square and a new approach is provided for the fast measurement of the K value by NIR spectroscopy.
出处
《红外》
CAS
2007年第8期37-40,共4页
Infrared
关键词
近红外
正交信号校正
聚氯乙烯
K值
near infrared
orthogonal signal correction
ploy （vinyl chloride）
K value
