样品选择结合分段直接校正法和偏最小二乘法用于近红外光谱分析模型传递研究

发布时间：2018-03-13 06:11

本文选题：光学　切入点：光谱分析　出处：《兵工学报》2016年01期 　论文类型：期刊论文

【摘要】：为了实现近红外光谱分析模型在不同仪器间的共享,对分析模型的传递方法进行了研究。以混胺燃料的分析模型为例,采用马氏距离法对转换集样品进行选择,采用分段直接校正(PDS)法结合偏最小二乘(PLS)法,对从仪器光谱进行校正。对转换集样品数、PLS回归的主因子数、PDS的窗口宽度进行选择,当转换集样品数为12、主因子数为1、窗口宽度为39时,转换集样品的光谱校正率达到96.5%.混胺分析模型传递后,预测集样品的标准偏差远远好于校正前的结果,与主仪器的非常接近。研究结果表明:采用马氏距离法选择转换集样品,并结合PDS-PLS算法计算转换矩阵,可有效扣除不同仪器间的光谱差异,提高混胺分析模型的传递精度,实现三乙胺、二甲苯胺、水分、密度的准确分析。
[Abstract]:In order to share the near infrared spectrum analysis model among different instruments, the transfer method of the analysis model was studied. Taking the analysis model of mixed amine fuel as an example, the Markov distance method was used to select the sample of the conversion set. In this paper, the method of piecewise direct correction (PDS) combined with partial least squares (PLS) method is used to correct the spectrum of the instrument. The window width of PDS is selected as the main factor of the conversion set sample number and PLS regression. When the sample number of conversion set is 12, the number of main factors is 1, and the window width is 39, the spectral correction rate of the sample is up to 96.5.The standard deviation of the predicted sample is much better than that of the sample before correction after the transfer of the mixed amine analysis model. The results show that using Markov distance method and PDS-PLS algorithm to calculate the conversion matrix can effectively deduct the spectral difference between different instruments and improve the transfer accuracy of the mixed amine analysis model. The accurate analysis of triethylamine, xylene amine, moisture and density is realized.
【作者单位】：海军航空工程学院飞行器工程系;
【分类号】：O657.33

【相似文献】