【作者】 覃旭松；

【导师】 戴连奎；

【作者基本信息】 浙江大学 ， 模式识别与智能系统， 2004， 硕士

【摘要】 在炼油过程以及石油产品销售中，通常需要对产品的关键品质比如汽油辛烷值等进行分析测定。传统的实验室分析方法不仅测定费用高，测量滞后大，而且操作和维护复杂。拉曼光谱分析技术是一种快速的无损分析技术，可用于对石油产品质量的离线和在线实时分析。为此，本文对拉曼光谱分析技术进行了深入研究，并开发了一种新型汽油辛烷值拉曼测定仪，具体包括以下几个内容： 1．通过阅读大量的中英文文献，对拉曼光谱分析技术的概念、原理及其在石油产品质量分析中的应用做了较为系统完整的阐述。 2．将小波变换应用于汽油拉曼光谱信号的噪声滤除。与常用的噪声滤除方法相比，小波变换能够较好地减弱荧光背景干扰和高频噪声，提高光谱数据的信噪比，显著提高了基于偏最小二乘(Partial Least Squares, PLS)方法所建立的汽油研究法辛烷值、马达法辛烷值预测模型的精度。 3．将最小二乘支持向量机(Least Squares Support Vector Machine, LS-SVM)算法应用于汽油拉曼光谱定量校正，并结合实际应用需要，提出了一种自适应建模方法。通过对一批汽油研究法辛烷值样本数据的仿真研究，结果表明LS-SVM模型比常用的PLS模型具有更好的预测效果与鲁棒性。 4．基于上述研究成果，并结合实际应用情况，开发研制了汽油辛烷值拉曼测定仪。针对某石油产品销售公司提供的一组成品汽油样本，该测定仪已显示出其良好的预测性能和较高的精度。更多还原

【Abstract】 Some properties of petroleum products such as gasoline octane number are necessary to be analyzed regularly. Traditional laboratory analysis methods are usually not suitable because of their cost, long delay, complication of operation and maintenance. Raman spectroscopy is a non-destructive and rapid analysis method, which is preferable to rapidly determine the properties of petroleum products offline or online. This thesis researched the application techniques of Raman spectroscopy and developed a new gasoline octane number analyzer using Raman spectroscopy. The main contributions of this thesis are as follows:1. Introduce the principles of Raman spectroscopy, and then review the application of Raman spectroscopy in petroleum products analysis.2. Apply wavelet transform in the pre-processing of gasoline Raman spectra in order to reduce the fluorescence backgrounds and to improve the signal-to-noise ratio of the spectra. Experimental results show that the application of wavelet transform obviously improves the prediction performance of Partial Least Squares (PLS) model for research octane number (RON) and motor octane number (MON), comparing with regular pretreatment methods.3. To overcome the disadvantages of linear calibration methods such as PLS, least squares support vector machine (LS-SVM) is introduced to the quantitative calibration of gasoline using Raman spectroscopy. Besides, a novel adaptive modeling technique is proposed in order to meet the practical need of rapid determination of gasoline octane number. For a set of commercial gasoline samples, the LS-SVM model obtains better performance than the PLS model on the prediction of RON.4. Based on the above research results, a gasoline octane number analyzer using Raman spectroscopy has been developed. For a set of gasoline samples, the analyzer performs satisfactorily on the prediction of RON and MON.更多还原