基于诊断比值的Logistic回归分析对中东原油的鉴别

发布时间：2018-06-10 15:57

本文选题：正构烷烃 + 诊断比值　；参考：《大连海事大学》2017年硕士论文

【摘要】：本文以7种中东原油和13种非中东原油为研究对象,采用GC-FID法对其进行检测分析,并探讨了风化对原油正构烷烃分布及诊断比值的影响。再通过相似性分析和聚类分析筛选建模参数,建立了 Logistic回归模型对中东地区原油进行了鉴别区分。中东原油的正构烷烃主要集中在n-C8～n-C17,以n-C13为分界点,n-C8～n-C13含量逐渐增加,在n-C13之后含量逐渐减少,总体分布趋势相似。非中东原油之间正构烷烃分布差异性较大,大部分非中东原油的正构烷烃主要集中在n-C13～n-C25,少数轻质组分含量较多。中东原油与非中东原油在诊断比值上存在着较大的差异。风化30天后,所有原油n-C13之前的正构烷烃组分都已损失完全,而诊断比值n-C17/Pr、n-C18/Ph和Pr/Ph基本上不受风化影响,仍可作为风化溢油鉴别的依据。以n-C17/Pr、n-C18/Ph、Pr/Ph值为特征向量分析原油之间的相似性。中东原油之间相似度较高,而非中东原油与中东原油之间相似度偏低,表明中东原油与非中东原油之间存在着显著的差异,两类油种可作为二分类事件加以分析建模。相对于正构烷烃,以诊断比值为特征变量能够更准确地将中东原油与非中东原油加以聚类区分,所以筛选出n-C17/Pr、n-C18/Ph、Pr/Ph作为建模参数。以诊断比值为参数建立的二元Logistic回归模型,其油种判别结果与预测目标匹配度高。实例验证结果显示,该模型对风化油样和文献数据中原油的判别准确率分别为100%、95%。表明油种Logistic模型不仅适用于未风化原油的鉴别,也适用于风化原油的鉴别,这为准确判别海上风化溢油种类提供了一个技术支持。
[Abstract]:In this paper, 7 kinds of Middle East crude oil and 13 kinds of non Middle East crude oil were studied by GC-FID method, and the influence of weathering on the distribution of normal alkanes and the diagnostic ratio of crude oil was discussed. Then the modeling parameters were screened by similarity analysis and cluster analysis, and the Logistic regression model was established to identify the crude oil in the Middle East region. The normal paraffin in Middle East crude oil is mainly concentrated in n-C8 to n-C17, with n-C13 as the demarcation point, the content of n-C8 to n-C13 is gradually increased, and the total distribution trend is similar after n-C13. The distribution of normal paraffin in non Middle East crude oil is very different, and the normal paraffin in most non Middle East crude oil is mainly concentrated in n-C13 to n-C25. There are large differences in the number of light components. There is a great difference in the diagnostic ratio between Middle East crude oil and non Middle East crude oil. After 30 days of weathering, all the normal paraffin components before n-C13 are completely lost, and the diagnostic ratio n-C17/Pr, n-C18/Ph and Pr/Ph are not affected by weathering, and can still be used as the basis for the identification of weathered spillage oil. N-C17/Pr, N-C18/Ph, Pr/Ph value is a characteristic vector analysis of the similarity between crude oil. The similarity degree between Middle East crude oil is high, but the similarity between Middle East crude oil and Middle East crude oil is low. It indicates that there is a significant difference between Middle East crude oil and non Middle East crude oil. The two kinds of oil can be analyzed and modeled as two classification events. The diagnostic ratio is a characteristic variable that can classify Middle East crude oil and non Middle East crude oil more accurately, so n-C17/Pr, n-C18/Ph and Pr/Ph are selected as modeling parameters. The two yuan Logistic regression model based on the diagnostic ratio is established, and its oil seed discrimination result is well matched with the prediction target. The example verification results show that the model is a model. The discriminant accuracy of the refined oil samples from the weathered oil samples and the literature data is 100%, respectively. 95%. shows that the Logistic model is not only applicable to the identification of the crude oil, but also for the identification of the weathered crude oil, which provides a technical support for the accurate discrimination of the types of the sea weathering spillage.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X55

【相似文献】